Leibniz Institut für Meereswissenschaften
The integration of microalgae photobioreactors in a recirculation system for low water discharge mariculture Dissertation zur Erlangung des Doktorgrades der Mathematisch Naturwissenschaftlichen Fakultät an der Christian Albrechts Universität zu Kiel vorgelegt von Nicole Kube
Kiel, 2006 Referentin: Prof. Dr. Karin Lochte Koreferent: Prof. Dr. Dr. h.c. Harald Rosenthal Tag der mündlichen Prüfung: Zum Druck genehmigt: Kiel, den Der Dekan
Foreword
The manuscripts included in this thesis are prepared for submission to peer reviewed journals as listed below: Wecker B., Kube N., Bischoff A.A., Waller U. (2006). MARE – Marine Artificial Recirculated Ecosystem: feasibility and modelling of a novel integrated recirculation system. (manuscript)
Kube N., Bischoff A.A., Wecker B., Waller U. Cultivation of microalgae using a continuous photobioreactor system based on dissolved nutrients of a recirculation system for low water discharge mariculture (manuscript)
Kube N. And Rosenthal H. Ozonation and foam fractionation used for the removal of bacteria and parti cles in a marine recirculation system for microalgae cultivation (manuscript) Kube N., Bischoff A.A., Blümel M., Wecker B., Waller U. MARE – Marine Artificial Recirulated Ecosystem II: Influence on the nitrogen cycle in a marine recirculation system with low water discharge by cultivat ing detritivorous organisms and phototrophic microalgae. (manuscript)
This thesis has been realised with the help of several collegues. The contributions in particular are listed below: Chapter 2: MARE I MARE was designed, constructed and daily maintained by Adrian A. Bischoff, Bert Wecker and Nicole Kube. Sampling and analyzing was done by Nicole Kube (daily maintenance of the recirculation system, fish biomass, dissolved nutrients, foam fractionation and supporting help for worm biomass), Adrian Bischoff (daily maintenance of the re circulation system, detritivorous tank sampling, fish biomass, dis solved nutrients, supporting help for foam fractionation) and Bert Wecker (macroalgae biomass, supporting maintenance of the recircula tion system). Bert Wecker developed the model and made the figures. Nicole Kube wrote the manuscript, supported by Adrian Bischoff. Dr. Martina Blümel and Dr. Uwe Waller reviewed the manuscript. Chapter 3: Photobioreactorsystem Nicole Kube designed the photobioreactor system, did the sampling and analyzing, supported by Adrian A. Bischoff and Bert Wecker. The manuscript was written by Nicole Kube, reviewed by Dr. Uwe Waller. Chapter 4: Foam fractionation Nicole Kube did the sampling and analyzing of the data. Nicole Kube wrote the manuscript, supported by Prof. Dr. Harald Rosenthal. Chapter 5: MARE II Nicole Kube and Adrian A. Bischoff did the sampling, analyzing and daily maintenance of the system. The manuscript was written by Nicole Kube, supported by Adrian A. Bischoff and Dr. Martina Blümel. Bert Wecker supported the modelling of the data. Uwe Waller reviewed the manuscript. Table of Contents
Chapter 1 ...... 5
1.1 Environmental impacts of open mariculture systems...... 7
1.2 Requirements of recirculation systems ...... 11 1.2.1 Feed uptake ...... 12 1.2.2 Biogeochemical cycles...... 12 1.2.3 Applications of biogeochemical cycles in aquaculture systems .... 16 1.2.4 Suspended and settable solids (Particles)...... 19 1.2.5 pH and alkalinity...... 20 1.2.6 Oxygen and CO 2 ...... 21 1.3 Technical recirculation system at IFM GEOMAR ...... 21
1.4 Recirculation systems with different trophic levels...... 23
1.5 Organisms ...... 26
1.6 Thesis outline ...... 29
1.7 References ...... 31
Chapter 2 ...... 39
2.1 Introduction...... 41
2.2 Material and Methods ...... 42 2.2.1 MARE System ...... 42 2.2.2 Measurements and Methods...... 44 2.2.3 Modelling...... 46 2.3 Results ...... 61 2.3.1 Feasibility of the MARE system...... 61 2.3.2 Modelling the nutrient budget...... 63 2.4 Discussion ...... 74 2.4.1 Feasibility of the MARE system...... 75 2.4.2 Nutrient recycling by integration of secondary organisms (Solieria, Nereis ) ...... 76 2.4.3 Modelling...... 81 2.5 Conclusions ...... 83
2.6 Acknowledgements...... 84
2.7 References ...... 84 Chapter 3 ...... 89
3.1 Introduction...... 91
3.2 Material and Methods ...... 93 3.2.1 Design of the continuous photobioreactor system ...... 93 3.2.2 Functional principle of the photobioreactors ...... 96 3.2.3 Algae ...... 100 3.2.4 Culture conditions...... 100 3.2.5 Sampling and analytical methods ...... 100 3.3 Results ...... 102 3.3.1 Feasibility of the photobioreactor system for algae cultivation ... 102 3.3.2 Specific growth rates and nutrient uptake rates of Nannochloropsis at different light intensities ...... 107 3.4 Discussion ...... 112 3.4.1 Applicability of the photobioreactor design...... 112 3.4.2 Nutritional value of microalgae ...... 113 3.4.3 Growth performance of Nannochloropsis spec. in continuous cultures...... 115 3.4.4 Filter efficiency of microalgae photobioreactors ...... 118 3.5 Conclusion...... 118
3.6 Acknowledgements...... 118
3.7 References ...... 119
Chapter 4 ...... 123
4.1 Introduction...... 125
4.2 Material and Methods ...... 127 4.2.1 System configuration...... 127 4.2.2 Sampling methods...... 128 4.3 Results ...... 132 4.3.1 Viable counts ...... 132 4.3.2 Quantitative and qualitative analysis ...... 134 4.3.3 Influence of ozone on efficiency of foam fractionation ...... 138 4.4 Discussion ...... 139
4.5 Conclusion...... 142
4.6 Acknowledgements...... 142
4.7 References ...... 143 Chapter 5 ...... 147
5.1 Introduction...... 149
5.2 Material and Methods ...... 150 5.2.1 Modifications of the recirculation system ...... 150 5.2.2 Measurements...... 152 5.3 Results ...... 154 5.3.1 Module fish ...... 154 5.3.2 Module detritivorous tank...... 155 5.3.3 Module microalgae bioreactors...... 159 5.4 Discussion ...... 161 5.4.1 Module fish tank ...... 161 5.4.2 Module detritivorous tank...... 162 5.4.3 Module microalgae bioreactors...... 162 5.4.4 Nitrogen cycle...... 163 5.4.5 General recommendations ...... 167 5.5 Acknowledgements...... 167
5.6 References ...... 168
Summary Summary The development of mariculture recirculation systems increasingly compre hends not only reprocessing of the water body but also an enhanced nutrient recycling by integration of secondary modules like macro and microalgae and detritivorous organisms. These modules are applied to utilize dissolved and particulate waste derived from fish cultivation in an environmental friendly manner. In this thesis a potential conceptual design of such a recirculation system is presented, developed by the mariculture group at IFM GEOMAR. The under lying principle of the recirculation system was based on that of an artificial ecosystem and combined several trophic levels (fish, macroalgae, microalgae, worms) (MARE = Marine Artificial Recirculated Ecosystem). Corresponding investigations showed that this type of recirculation system becomes feasi ble, if the dimensions of the secondary modules are adapted to the biotic and abiotic culture requirements of the target species. The results of the experi ments provided substantiated knowledge regarding nutrient cycles within the recirculation system, which could be described by a numeric model. The main goal of this thesis was the development of a continuous photobio reactor system for cultivation of Nannochloropsis spec. based on dissolved nutrients derived from a marine recirculation system. From the results it be came evident, that it is possible to gain additional biomass and to return al most 100 percent of the derived water back to the main water cycle. The data also revealed that the biofilter capacity of photobioreactor system remains strictly limited to a low level. Furthermore, the efficacy of ozone and foam fractionation was investigated regarding removal of bacteria and particles from effluents of a marine recir culation system. Interestingly, it could be proved by using a specific staining method that most of the living bacteria are attached to particles with the size of up to 50 m which protects these organisms efficiently from being killed
1 Summary by treatment with ozone. Thus, reduction of total bacteria numbers could be achieved by the removal of the particles by foam fractionation. A second trial of the MARE system with fish, worms and microalgae re vealed, that recirculation systems with several trophic levels require a proper management. If worms are not removed from the recirculation system before natural spawning, nitrogen cycle of the recirculation system can be nega tively influenced. This leads to an impaired growth of microalgae and unsta ble culture conditions for fish. Taken together, this work demonstrated for the first time, that a marine eco system could be artificially mimicked by installation of a cycle of nutrients over several trophic levels. Hence, these results may contribute to reduce environmental impacts of maricultures in the future.
2 Summary Zusammenfassung Die Entwicklung von Marikultur Kreislaufsystemen beinhaltet zunehmend nicht nur die vollständige Wiederaufbereitung von Wasser, sondern auch ein erweitertes Nährstoffrecycling durch die Integration von sekundären Modu len wie Makro und Mikroalgen und detritivoren Organismen. Diese Module werden benutzt, um die anfallenden gelösten und partikulären Abfallstoffe der Fischproduktion umweltgerecht weiterzuverwerten.
In dieser Dissertation wird ein mögliches Konzept einer solchen Anlage vor gestellt, die von der Marikulturgruppe am IFM GEOMAR entwickelt wurde. Das Prinzip der Anlage beruhte auf einem künstlichen Ökosystem und kom binierte verschiedene trophische Stufen (Fische, Makroalgen, Mikroalgen, Würmer) (MARE = Marine Artificial Recirculated Ecosystem). Die Untersu chungen haben gezeigt, dass die Machbarkeit einer solchen Anlage gegeben ist, wenn die Dimensionierungen der sekundären Module an die biotischen und abiotischen Bedingungen der Zielart angepasst sind. Die Ergebnisse der Experimente lieferten fundierte Kenntnisse in Bezug auf die Nährstoffkreis läufe innerhalb des Kreislaufsystems, die durch ein numerisches Modell be schrieben werden konnten. Zentrales Thema dieser Dissertation ist die Entwicklung eines kontinuierli chen Photobioreaktorsystems zur Kultivierung von Nannochloropsis sp. auf der Basis gelöster Nährstoffe aus einem marinen Kreislaufsystem. Die Ver suche haben gezeigt, dass die Gewinnung von zusätzlicher Mikroalgenbio masse grundsätzlich möglich ist und das gereinigte Wasser zu fast 100% wieder an den Hauptkreislauf zurückgegeben werden kann. Die Ergebnisse wiesen aber auch darauf hin, dass das entwickelte Photobioreactorsystem nur eine beschränkte Biofilterleistung erbringen kann. Des weiteren wurde die Effizienz von Ozon und Abschäumung im Hinblick auf die Entfernung von Bakterien und Partikeln aus dem Abwasser einer marinen Kreislaufanlage untersucht. Interessanterweise konnte durch die Anwendung einer spezifischen Färbemethode nachgewiesen werden, dass die
3 Summary meisten lebenden Bakterien an Partikeln der Größe bis 50 m haften, wo durch sie den Behandlungsprozess mit Ozon überstehen können. Dennoch war es möglich, die Gesamtbakterienzahl durch die Entfernung der Partikel mittels Abschäumung zu reduzieren. In einem zweiten Versuchslauf der MARE Anlage mit Fischen, Würmern und Mikroalgen wurde deutlich, dass eine Kreislaufanlage mit verschiedenen trophischen Ebenen eines korrekten Managements bedarf. Werden Würmer vor ihrer natürlichen Reproduktion nicht aus dem System entfernt, kommt es zu einer Anreicherung von organischem Material im Kreislaufsystem, das den Stickstoffkreislauf des Kreislaufsystem nachhaltig negativ beeinflussen kann. Das führte zu einer Limitierung des Mikroalgenwachstums und insta bilen Kulturbedingungen für die Fische. Zusammenfassend wurde mit dieser Arbeit erstmalig dargelegt, dass durch den Aufbau eines erweiterten Nährstoffkreislaufs über verschiedene trophi sche Stufen ein marines Ökosystem artifiziell nachgebildet werden kann. Daher können die Ergebnisse einen Beitrag leisten, in Zukunft die Umwelt einflüsse von Marikulturen zu reduzieren.
4
Chapter 1 General introduction and thesis outline
5
6 Chapter 1 1.1 Environmental impacts of open mariculture sys- tems The gradual decline in supplies of seafood from the ocean is one of the great est challenges of the seafood industry nowadays. For the past decade the worldwide capture fishery industry has been stagnant due to overfishing. No short term recovery from the current situation can be expected in the future. However, consumer demand for high quality seafood at reasonable prices is increasing. Often a year round availability of fresh seafood is expected. Hence these factors directly lead to a rapid expansion of aquaculture: this sector is showing an above average economic growth. Estimations of a total aquaculture output of 62 millions tons per year in 2025 are published (FAO 2002), representing a duplication of today´s production (Fig. 1). The Food and Agriculture Organization of the United Nations (FAO) defines aquacul ture as the farming of aquatic organisms including fish, crustaceans, mol luscs and aquatic plants. Aquaculture can thus be very diverse. To date, commercial aquaculture is dominated by cultivation of kelp, carp, oysters and tiger prawns (FAO, 2005). The majority of organisms are cultivated within freshwater aquaculture systems (57%, Davenport, 2003). Until now, marine aquaculture (mariculture) accounts for a smaller proportion (43%) but is of growing importance because of the stagnating fishery industry. The application of a marine recirculation system for cultivation of Gilthead seabream (Sparus aurata ) will therefore be the focus of this work.
200 Capture Fisheries 180 Aquaculture 160
140
120
100
80
60 Fish Supply [Mio. MT] Supply [Mio. Fish
40
20
0 1976 1980 1990 1997 2000 2010 2020 2030 Year Fig. 1 Global fish supply, data presented from 1976 till 2000 as recorded data, 2010 to 2030 as projection (FAO, 2002). 7 Chapter 1
In Europe, mariculture is dominated by the cultivation of Atlantic salmon (Salmo salar ), Gilthead seabream ( Sparus aurata ) and Sea bass ( Dicentrachus labrax ) (Cross, 2003). In total, 80.000 tons of seabream were cultivated in 2002 mainly in the Mediterranean region. Greece was the major contributor to seabream mariculture (49%), followed by Turkey (15%), Spain (14%) and Italy (6%) (FAO, 2005) (Fig. 2b). The intensive culture of carnivorous fish species like salmon causes greater environmental impacts than extensive cultivation methods for herbivorous species like carp, oysters and mussels. a) b)
Fig. 2a ) Net cages in Greece (Photo by H. Thetmeyer) b) Annual production of seabream in the EU region by mariculture (FAO, 2005). The production was stagnant in recent years because enlargement of the cultivation units was not possible due to regulations of environmental impacts and conflicts with tourism (Cross, 2003) In Europe, mariculture is mainly organized using cages along the coastline (Fig. 2a). Breeding in these systems is simple and economical, but the cages are directly connected with the environment. In most cases the cages are in stalled in wind protected areas, e.g. bays or fjords. The high load of organic (faeces and feed remains) and inorganic waste (fish excretory products) are directly leading to eutrophication in the surrounding environment (primary effects, Fig. 3) (Ackefors and Enell, 1994; Beveridge et al ., 1991). Addition ally, trace elements and micronutrients (e.g. vitamins from the feed) or re sidual pharmaceuticals can be found in the environment surrounding the cages. Furthermore these open systems are easily affected by economical damages generated by the transfer of diseases, parasites and toxic sub stances (Braaten et al. , 1983; 1992; Weston 1991).
8 Chapter 1
Fig. 3 Inputs and outputs from a fish cage culture system (from Davenport et al., 2003): Feed intake causes an increased demand of oxygen and releases a lipid film at the water surface. Due to the fish metabolism CO 2 and ammonia nitrogen are excreted. Particulate waste, consisting of faeces material and uneaten fish feed, deposit at the sediment around the net cages. The organic load causes an oxy gen drain due to bacteria activity. When the demand of oxygen exceeds the oxygen diffusion rate from overlying waters, sediment becomes anaerobic. This has on one hand a strong influence of the ben thos ecosystem due to decreasing oxygen tension. Also a range of microbial processes can follow up: nitrification (ammonium oxidation to nitrite and nitrite oxidation to nitrate) is not taking place, deni trification (producing dinitrogen from nitrate) is competing with nitrate reduction (producing again toxic ammonia from nitrate). Also toxic hydrogen sulfide can be produced by sulfate reduction and under most reducing conditions also methane (via methanogenesis). A second problem is the interaction with wild fish populations: unplanned releases of farmed fish (e.g. by destroyed nets) and/or gametes or fertile eggs can interfere with the genetic pool of wild stocks, resulting in interbreeding and reduced wild stock fitness and fertility. Furthermore, artificial structures used in mariculture affect local ecosystems: nets, anchors, mooring lines etc. can cause entanglement of the marine wildlife and provide substrates fa vouring fouling. Other species of fish, birds, marine mammals and also rep tiles are attracted by the mass of cultivated fish (Nash et al. , 2005). Very often mariculture activities are in conflict with other human interests. Competition for land and water are often the main targets. Especially con flicts with tourism limits growth of mariculture (Cross, 2003). Using coast lines for mariculture additionally destroys original biotopes and ruins fish breeding and nursery areas. Besides, local ecosystems (e.g. coral reefs) are affected by increased erosion, which secondarily effects the local fisheries. Modifications of the natural food web due to the fishing of wild juvenile fish
9 Chapter 1 for breeding/mast, the consumption of planktonic organisms due to large scale bivalve farming and increased fishing pressure on small pelagic fish populations for aquaculture feed (Nash et al. , 2005) have to be considered as long term effects resulting from this form of aquaculture. Therefore, a sustainable development of the aquaculture sector is required “to ensure the attainment and continued satisfaction of human needs for present and future generations. Such sustainable development (in the agri culture, forestry and fisheries sector) conserves land, water, plant and ani mal genetic resources, is environmentally non degrading, technically appro priate, economically viable and socially acceptable” (FAO, 1988). In particu lar, a sustainable management in the aquaculture sector comprises efficient use of land, water, energy and animal nutrition with simultaneous consid eration of bearing capacity of the aquatic environment and adjacent ecosys tems. In recent years great efforts have been made to reduce the environmental impacts of net cages: improvements in siting, design, technology, and man agement at farm level including improved feeding with a lower waste dis charge, better fish health management including disease and stock control at individual farm and sector level and the investigation of social and eco nomical aspects are only a few issues (GESAMP, 2001). Nevertheless, net cages are still open systems with negative impacts on the natural surrounding environment. The development of recirculating systems with partly or complete reuse of water (Losordo, 1999; Waller, 2000) has been promoted within the past decades, especially due to improvements in system technology (Rosenthal and Grimaldi, 1990; Losordo et al. , 1999; Summerfelt, 2002; Waller et al. , 2003). These closed cultivation systems can ensure the constant market availability of the cultured marine organisms by providing controlled system conditions. They can be a potential alternative to conventional production systems when failure free operation and profitability are assured. To date, intensive re
10 Chapter 1 search is carried out dealing with closed recirculation systems. Further ap plications for recirculation systems are ornamental/tropical fish culture, ma ture and brood stock culture, fry and fingerling production and niche mar kets for high price food fish. 1.2 Requirements of recirculation systems Within a recirculation system a suitable environment and good water quality has to be maintained. The recirculating water (process water) is subject to several treatment processes leading to an elimination of harmful substances or enriched nutrients (anoxic conditions are favoured by nutrient enrich ment). Primary conditions to ensure survival of the fish are: sufficient con centration of dissolved oxygen (>5mg/L or 50% DOT), low concentrations of undissociated ammonia nitrogen (<1mg/L), nitrite nitrogen (< 1mg/L) and carbon dioxide (<0.2 ppm) (Losordo et al. , 1998). Additionally, pH values are required to be in a range from 6 to 9 (optimum for seawater: 7 to 8, Losordo, 1999). If less than 10% d 1 of the total system volume are exchanged, a recircula tion system is called closed recirculation system (EIFAC, 1986), although modern systems can reach << 1% d 1. water exchange rate. A system meet ing these requirements has been developed at IFM GEOMAR (Waller et al. , 2003). Recirculating systems are biologically complex systems and need a sophisti cated planning. The critical parameters vary according to biological, chemi cal and physical interactions between the different system components, es pecially with intensive fish rearing (0.04 kg L 1 , Losordo et al. 1998). Accord ing to this study, an 80 L commercial home aquarium can be stocked with 5 kilograms of fish. The key to an efficient mariculture in a closed recirculating system is the use of high effective water treatment components and a good daily maintenance. In poorly managed systems the main problem is insufficient water quality
11 Chapter 1 because of inappropriate components and failures. This leads to stress, dis eases and high mortalities among the fishes. Uniform flow rates of water and air/oxygen, fixed water levels and continuous operation are of primary im portance. Especially the amount of exchanged water is decisive: the lower the amount of discharged water, the more important are detailed knowledge of and experience with recirculating systems. An introduction dealing with the critical parameters for maintenance of a recirculating system is given in the following chapter. 1.2.1 Feed uptake It is necessary to raise a fish from 5g body weight to market size of approxi mately 350 400g in about one year to generate a profit. To ensure such growth, fish are fed using high protein pellets at rates from 0.8 to 10 percent of body weight depending on size and species. The feed uptake has an enor mous influence on the environmental conditions of the recirculation system. The first effect appears directly after the feeding: due to fish digestion, the respiration rate increases dramatically associated by a decrease of dissolved oxygen concentrations to critical levels. Especially after intensive feeding (2 or 3 portions of the total daily feed) (Masser et al., 1999) this effect can be observed. A longer feeding period up to 15 hours e.g. by installation of auto matic feeders is therefore preferred. There are also secondary effects of in tensive feeding. The fish do not use the whole content of nitrogen and phos phorus of the feed for biomass production: for example max. 30% of the total nitrogen of the feed remains in the fish body (Krom et al., 1985; Krom and Neori, 1989; Hall et al., 1992; Lupatsch and Kissil, 1998; Hargreaves, 1998). Thus, 70% of the feed nitrogen content is excreted as organic waste: it can be classified into settling and suspended solids and excreted dissolved nutri ents. The main challenge of a successful maintenance of a recirculation sys tem is the control of the high organic waste. A constant feeding is therefore considered as decisive for the system management. 1.2.2 Biogeochemical cycles Two biogeochemical cycles are of major importance for recirculation systems: the nitrogen cycle and the phosphorus cycle. The nutrition of the fish is en
12 Chapter 1 riched in nitrogen as well as in phosphorus. Feed remains are subject to bio geochemical recycling within a recirculation system as well as the excretion products of the cultivated organisms. So it is important to understand the impact of the biogeochemical nutrient cycling in natural as well as in artifi cial systems. In the following chapters, the nitrogen as well as the phospho rus cycle are firstly presented in natural systems; then implications and ap plications for aquaculture systems are described.
Nitrogen cycle The nitrogen cycle is a complex biogeochemical cycle in natural systems (Fig. 4). The nitrogen cycle is also of special interest in aquacultural systems in order to observe and control the bacterial nitrogen metabolism because toxic intermediate products (e.g. ammonia, nitrite) can form and be enriched in the circulation water. There are two major processes of nitrogen conversion: nitrification and denitrification.
organic matter
+ NH 4 Nitrification N2-fixation
NH 2OH AN AM M OX Ammonification
NO - 2 NO N2O N2 Assimilatory nitrate reduction
Denitrification
- NO 3 Fig. 4 : Nitrogen cycle. Aerobic processes are marked in blue arrows, anaerobic processes are marked in black arrows. The ANAMMOX process is marked with red boxes.
13 Chapter 1
Nitrification The oxidation of organic or inorganic nitrogen compounds is called nitrifica tion. It is an aerobic process resulting in the formation of nitrate (NO3 2 ). Ni trification is performed by nitrifying bacteria (nitrifiers) being widely distrib uted in the terrestric and marine realm. Nitrifiers can be divided into ammo nia oxidising bacteria (AOB, e.g. Nitrosomonas, Nitrosococcus, Nitrosospira, Nitrosolobus, Nitrosovibrio ) converting ammonia to nitrite and nitrite oxidising bacteria (NOB, e.g. Nitrobacter, Nitrospina, Nitrococcus, Nitrospira ) converting nitrite to nitrate (Bock et al., 1986; Bothe, 2000). Nitrification can be performed both by autotrophic and heterotrophic bacteria. Autotrophic nitrifiers are able to combine nitrification with denitrification converting the
NO 2 produced by ammonia oxidation to N 2 (nitrifier denitrification, Wrage et al. , 2001). Heterotrophic nitrifiers gain no energy performing nitrification (Schmidt et al. , 1999) and therefore may have selective advantages under suboxic conditions. In aquaculture systems the nitrification process is of major importance be cause of the removal of toxic compounds. When organic material is degraded by bacteria, ammonia is formed. Ammonia is also excreted by the fish. Two compounds of ammonia are present: dissociated and undissociated ammo nia. The undissociated ammonia can rapidly reach toxic concentrations for fish and therefore needs to be removed. Therefore, nitrification is an impor tant process. This also accounts for nitrite being toxic for fish in comparably low concentrations. In contrast, nitrate, the final product of the nitrification process is not known to be a toxic nitrogen compound for fish, even in ele vated concentrations. However, nitrification can also be performed in the re verse direction, producing ammonia by nitrate reduction. Reduction equiva lents from fermentation processes are oxidised during nitrite reduction re sulting in the formation of ammonia (nitrate ammonification). Additionally, nitrate reduction can be used for an assimilatory process (assimilatory ni trate reduction, Schlegel, 1992). Nitrate ammonification is performed by fac ultative anaerobic bacteria and occurs during anoxic conditions. In contrast, the majority of bacteria carries the enzymes for assimilatory nitrate reduc tion. This process is performed during oxic as well as anoxic conditions
14 Chapter 1 when nitrate is the only available nitrogen source. Assimilatory nitrate re duction is inhibited by elevated ammonia concentrations (Schlegel, 1992). Another recently discovered process, the anaerobic ammonia oxidation (ANAMMOX; Strous, 1999), removes ammonia via the formation of nitrite forming N 2. However, the ANAMMOX process is an anaerobic process. An aerobic conditions are not desired to develop in aquaculture systems and therefore the ANAMMOX process is not likely to play a major role. Due to the high particle load in a recirculation system (fish excretions), suboxic/ anoxic conditions are likely to occur at the particle surfaces and the ANAMMOX process therefore may contribute to the removal of ammonia.
Denitrification
The stepwise reduction of nitrate via nitrite (NO 2 ), nitric oxide (NO) and ni trous oxide (N 2O) is called denitrification. The final product of the denitrifica tion process is atmospheric nitrogen (N 2). Pseudomonas sp., Bacillus sp., Thiobacillus sp. and Propionibacterium sp. are known to be denitrifying bac terial genera (Zumft, 1997; Wrage et al. , 2001). Also among the Archaea , genera possessing the enzymes for denitrification have been detected (e.g. Haloarcula marismortui, Zumft 1997). The majority of denitrifying bacteria are aerobic heterotrophic organisms using nitrate (nitrite, nitric oxide, ni trous oxide) as terminal electron acceptor molecules in the respiratory chain. Denitrification occurs when suboxic/anoxic conditions are prevailing, be cause the enzymes responsible for this process (e.g. nitrate reductase, nitrite reductase) are sensitive to elevated oxygen concentrations. Like nitrifiers, denitrifying bacteria are widely distributed in the terrestric as well as in the marine realm. Factors affecting denitrification rates are: nitrate concentra tion, concentration of electron donators (e.g. organic carbon compounds, re duced sulphur compounds, hydrogen) and the presence/absence of oxygen being considered as the decisive criterion (Tiedje 1988). Recently, aerobic denitrification was also observed for several bacterial species. The oxygen concentration tolerated by these bacteria is varying (Zumft 1997, Wrage 2001). In aquaculture systems, denitrification is restricted to special units where suboxic/anoxic conditions can be maintained (Chapter 3,5).
15 Chapter 1
Phosphorus cycle Compared to the nitrogen cycle, the biogeochemical cycling of phosphorus is rather simple. Inorganic phosphate is mainly brought into the marine realm by weathering of rocks. Inorganic phosphate often is a limiting factor for pri mary production in the oceans and is easily integrated into phytoplankton biomass by primary production. The phosphorus is then transferred biologi cally via the classical food chain to higher trophic levels. Particulate and dis solved phosphate is excreted by larger organisms of all trophic levels and particulate phosphate is produced during decay of organic matter. The detri tus sinks to the seafloor and in geological timescales, is fixed in rock mate rial. 1.2.3 Applications of biogeochemical cycles in aquaculture systems
Nitrogen cycle Total ammonia nitrogen (TAN) is a by product of the protein metabolism of fish (Bone and Marshall, 1995). TAN is excreted by the fish via the gills as digestion product and is also produced by bacteria converting organic waste into dissolved nutrients. TAN consists of undissociated ammonia (NH 3) and dissociated ammonia (NH 4+). The undissociated form is extremely toxic to most fish species. The relative proportions of undissociated and dissociated ammonia are strongly depending on pH and temperature and also, to a mi nor extent, on the salinity of the system water (Trussel 1972, Bower 1978). At pH 7.0 the majority of ammonia occurs in the dissociated form, but at pH 8.5 more than 30 percent of the TAN is detected as NH 3. Thus, it is of major importance to keep TAN concentrations as low as possible coinciding with the maintenance of pH values <8.5. In general, the concentration of dis sociated ammonia nitrogen should not reach more than 0.05mg/L (Losordo, 1998), although lethal concentrations can vary from species to species. Only few studies have been conducted examining the influence of sub lethal TAN concentrations on fish, but effects on growth and immune defence, skin damages and a reduced fertility was observed (Colt and Armstrong, 1979; Russo and Thruston, 1991; Hargreaves, 1998).
16 Chapter 1
Thus, it is one of the primary objectives during maintenance of a recircula tion system to avoid TAN accumulation and to eliminate this component from the system as soon as possible. To achieve TAN reduction/elimination, biological filtration is the preferred treatment step in aquaculture. There are several forms of biofilters, e.g. fluidized bed filters, mixed bed filters or trick ling filters (Losordo et al. , 1999, Fig. 5a). Within a biofilter a suitable sub strate (e.g. plastic pellets, Fig. 5c) providing an enlarged surface for attach ment especially for nitrifying bacteria (e.g. Nitrosomonas sp.) is supplied.
The first product of the nitrification process is nitrite (NO 2 ) being also a toxic component for the fish. Nitrite ions are taken up into the fish by the chloride cells of the gills, oxidising the Fe 2+ ion in the haemoglobin molecule (Boyd and Tucker, 1998). Methaemoglobin as the resulting product, is not able to reversibly bind oxygen (Colt and Armstrong, 1979; Russo and Thurston, 1991; Hargreaves, 1998; Kioussis et al. , 2000) being crucial for respiration. Therefore, with increasing nitrite concentrations in the water and hence in the blood circulation system the oxygen binding capacity of the blood de creases. Fish are forced to elevate the ventilation rate of the gills. Although the oxygen concentration of the water is sufficient for fish survival, suffoca tion is possible due to elevated nitrite concentrations. Toxicity of nitrite does not only depend on the fish species, but also on the pH and concentration of Cl and Ca 2+ ions in the water (Colt and Armstrong, 1979; Kioussis et al. , 2000). The recommended concentrations of Nitrite N for intensive aquacul ture is <0,1mgL 1 NO 2 N in freshwater and <1mg L 1 NO 2 N in seawater (Wickins, 1980). Assuming stable conditions in a recirculation system, nitrite is rapidly con verted to the nontoxic nitrate (NO 32 ). In biofilters used in aquaculture sys tems also bacterial species (e.g. Nitrobacter sp.) converting either nitrite or both ammonia and nitrite to nitrate are present, resulting in a complete re moval of the toxic nitrogen compounds by the process of nitrification (Rhein heimer et al. , 1988; Hagopian and Riley, 1998). High concentrations of ni trate are not supposed to be of major importance for fish survival, aquatic species can tolerate extremely high levels (>200mg/L). Therefore, only few
17 Chapter 1 studies are published discussing toxic levels and effects of nitrate (Russo and Thurston, 1991). However, there are indications of effects on the osmo regulation and oxygen transport in the blood as reaction to very high nitrate concentrations (Colt and Armstrong, 1979; Kioussis et al. , 2000). Within a completely closed recirculation system, as it is operated at the IFM GEOMAR (see Subsection 1.3), nitrate is accumulating in the system. The installation of a denitrification reactor can help to control the nitrate concen tration (Fig. 5b). In the denitrification reactor, anoxic conditions are main tained due to low flow rates. Additionally, methanol is added as a C source for the bacterial metabolism according to the redox potential in the reactor (Wecker, 2002). a) b) c)
Fig. 5.a) Contiwash biofilter with internal counter flow principle. The arrows show the flow direc tion of the pellets (Sander). b) Denitrification reactor, working analogous c) Plastic pellets provided as substrate for bacteria. They are used both in the biofilter and the denitrification reactor.
Phosphorus cycle The nutrition of the fish is enriched in phosphate. Around 70 to 90% of the phosphate provided with the feed is released to the system environment as
PO 43 . A proportion of approximately 20% thereof is released as dissolved phosphate, whereas the major proportion (80%) is excreted as particulate matter (Bodvin et al. , 1996). Phosphate is a nontoxic compound and high concentrations can be tolerated by the fish. In a recirculation system, accu mulation of phosphate can easily be avoided by proper water treatment, be cause most of the phosphate can be discharged with the organic solids or is 18 Chapter 1 used in bacterial metabolism during denitrification. (Barak and van Rijn, 2000). If the accumulation exceeds concentrations of more than 100 mg L 1 a water exchange can be considered, but the integration of phototrophic or ganisms can help to control the phosphate levels. 1.2.4 Suspended and settable solids (Particles) Especially the suspended (particles <50 m) and settable solids (particles >50 m, fish faeces) need to be removed rapidly from the system (Cripps and Bergheim, 2000). The fish faeces alloy the water quality: due to bacterial ac tivity and chemico physical conditions leaching (= extraction of nutrients) can be observed. Decomposition of remaining solids results especially in an increase of ammonia nitrogen concentrations and a decrease in dissolved oxygen tension (DOT) in the system (Welch and Lindell, 1992). Just like dur ing fish cultivation in an open system using net cages toxic compounds (e.g. hydrogen sulfide) can accumulate. The removal of particles is very important to control bacterial abundances in the recirculation water, because the ma jority of living bacteria can be found attached to particles (Kube & Rosenthal 2006, Chapter 4; Braaten, 1986; Litveld and Cripps, 1992). Assuming undisturbed water conditions (no flow), larger particles (settable solids) will settle down soon after release. These particles can be removed using a sedimentation tank, a mechanical filtration or a swirl separator. However, the suspended solids (smaller particle sizes) will not settle down and can cause turbidity. When these particles are not efficiently removed, the resulting turbidity can cause stress and irritations of the fish gills and consequently influence fish health (Rosenthal et al. , 1982). The use of foam fractionators is a proper and economical way for removing suspended sol ids. An air/ozone mixture is discharged into a closed cylinder at the bottom of the cylinder. The water is let in following the counter flow principle (Fig. 6). The bubbles rise through the water column, their active surface binds proteins and particles, forming foam at the top of the cylinder. The foam is then collected in a waste tank. The proportion of ozone supports the foam ing process: ozonation – through electrostatic loading and polarization of the hydrophobic hydrophilic ends foster settling characteristics of sus
19 Chapter 1 pended solids while also assisting in forming aggregates that attach to wa ter air interfaces, producing a stable foam which can be removed by counter current stripping. Ozone is able to break up organic compounds, long chain molecules and especially lipids forming a thin film on the water surface after feeding. Fig. 6 A Fresh Skim 200 (Sander) foam fractionator: water is pumped in following the counter flow principle. Through the airstone an air/ozone mixture is let in. A surface active layer is gener ated around the discharged bubbles. Proteins are accumu lating with their hydrophobic end to the bubble. The hydro philic end is catching particles esp. of organic origin. Due to the buoyancy of the bubbles and the following foam forma tion all particles are removed and collected in the upper foam collector. 1.2.5 pH and alkalinity The pH (concentration of H + ions) of water does effect numerous other water parameters (e.g. ammonia nitrogen). It also influences conversion rates of other biological and chemical processes. Thus, pH must be monitored to en sure optimal conditions for fish growth in recirculation systems (Losordo, 1998). The optimal pH range for fish health is 5 9 with an optimum from pH 7 to 8. Inadequate pH concentrations also influence metabolic rates of other important organisms in the system, e.g. below pH 7 the activity of nitrifying bacteria is reduced (Bischoff and Kube, unpubl. data) Alkalinity is defined as the capacity of water for acidity neutralization. Bicar bonate (HCO 3 ) and carbonate (CO 32 ) are the predominant compounds de termining alkalinity in seawater. The higher the alkalinity, the higher is the buffering capacity of the water against pH change. pH values can fluctuate
20 Chapter 1 or cycle daily due to respiration, whereas alkalinity is relatively stable but also alkalinity can change over longer time periods (Wurts et al. , 1992). Nitrification is an acid producing process; the transfer of ammonia nitrogen to nitrate nitrogen produces H + ions. Due to the reactions with hydroxide ions (OH ), carbonate and bicarbonate, alkalinity and pH values decrease. In order to maintain alkalinity and pH values, lime and sodium bicarbonate
(NaHCO 3) can be added to the water.
1.2.6 Oxygen and CO 2 The concentration of dissolved oxygen (DO) is the main limiting factor for the fish carrying capacity of a recirculation system. The overall rate of oxygen consumption in a recirculation system is determined by the respiration rate of the fish, the oxygen demand of bacteria consuming organic waste and food remains and the oxygen demand in the nitrifying biofilters (BOD – Biological oxygen demand) (Losordo, 1998). DO values should be above 60% of satura tion (DOT). During intensified cultivation and increasing feeding rates additional aera tion of the system is necessary. Required concentrations of dissolved oxygen can be maintained through continuous aeration either with atmospheric air or pure gaseous oxygen (Losordo, 1998). As mentioned before (subsession 1.2.3), the proper removal of solids can drastically reduce the oxygen de mand of the system. 1.3 Technical recirculation system at IFM-GEOMAR Modern recirculation systems allow the intensive culture of marine organ isms at almost optimal living conditions und minimized water exchange (Liao and May 1974; Otte and Rosenthal 1979, Bovendeur et al. , 1987, Blanche ton 2000, Waller et al. , 2001). One of these systems is a technical recirculation system installed at the Fishery Biology Department of IFM GEOMAR. Considering the special re
21 Chapter 1 quirements of such a system discussed in the preceding chapters the follow ing components were installed (Fig. 6): in two circular fishtanks Gilthead Seabream ( Sparus aurata ) was cultivated (1). Solid waste was removed by a two step separation process: larger particles sedimented in the swirl separa tor (2, Fig. 7b,c) and particles <50 m were removed with a foam fractionator Helgoland 700 with ozone addition (4, Fig. 7a). Ammonia nitrogen and nitrite was converted to nitrate in a Contiwash biofilter, the nitrate was removed by a denitrification bioreactor (5). Online measurements and automatically regulations of oxidation reduction potential (ORP), pH and oxygen levels were monitored with a control module (KM 2000). To maintain pH a CaO dosage unit was installed at the swirl separator (Fig. 7c).
4
1 1 6 3 5
7 2
Fig. 6 Schematic drawing of the technical recirculation system at IFM GEOMAR. Arrows show the flow direction of water. (1) fish tanks with aeration, (2) Swirl separator, (3) Contiwash Biofilter, (4) Foam fractionator, (5) denitrification, (6) Airlift, double triangle = tap
22 Chapter 1 a) b) c)
Fig. 7a ) Foam frac tionator (left) with vessel for rinsing water (right) b) on top view of a swirl separator with the central black up per outflow. Water rises with an eddy current from bottom inlet meanwhile larger particles can settle down c) Biofilter (left) and swirl separator (right) with CaO dos age device (middle) Water is aerated in the tanks using compressed air, but the foam fractiona tor and the biofilter water are enriched with oxygen as well. Due to effective treatment steps water discharge rates <<1% per day could be achieved (Waller et al., 2003b). Based on this knowledge and research a commercial scale recirculation system with 120m³ water volume and 10 tons of fish bio mass production was set up near Hanover (PISA – PolyIntegrated Seawater Aquaculture) (Waller et al. , 2005). 1.4 Recirculation systems with different trophic levels To date, in closed recirculation systems waste produced by fish was either removed or nontoxic nutrients, e.g. nitrate (resulting from biofiltration) and phosphate (leaching of solids) were accumulating. The daily amount of waste is very high: results from different aquacultural cultivation systems showed that only 20 30% of the nitrogen from the feed are used by the fish for bio mass synthesis (Krom et al. , 1985; Krom and Neori, 1989; Hall et al., 1992; Lupatsch and Kissil, 1998; Hargreaves, 1998). Utilization of phosphate is even lower: published values range from 10 to 30% of feed uptake (Krom et al., 1985; Krom and Neori, 1989; Barak and van Rij, 2000; Lupatsch and
23 Chapter 1
Kissil, 1998) in turn meaning, that 70 90% are excreted either in dissolved or in particulate form. Until recently, recirculation systems were designed to recycle the water only. Important nutrients like nitrogen, phosphorus or carbon compounds are therefore eliminated either by biological conversion or due to water exchange (Losordo et al. , 1999; Waller et al. , 2005). Thus, these valuable organic com pounds are lost for the system. In modern aquaculture systems a compre hensive nutrient recycling should be maintained in order to reduce the waste produced by the system. Future production systems without any nutrient and energy recycling are not supposed to be economically and ecologically successful (Troell et al. 2003). Integration in terms of aquaculture is defined as the controlled cultivation of aquatic organisms of different trophic levels and one of the key prerequisites for fulfilling the demands for sustainability. Waste from fish cultivation can be regarded as „new resources“ (Chamber lain and Rosenthal, 1995). This topic increasingly attracts international at tention. Due to the better utilization of these resources by joining different trophic levels the profitability of recirculation systems can be enhanced. Waste from the production of fish is used to create high quality biomass. Thus, the environmental impacts of a recirculation system can be reduced and resources can be used more efficiently (Asgard et al. , 1999; Schneider et al. , 2005). The additional biomass supports the economical diversification and the benefit per cultivation unit (Chopin et al. , 2001). However, the idea of integration is not new. In the aquaculture sector dealing with freshwater and brackish systems Asian countries have been practising integrated aquaculture for centuries (Chopin et al. , 2001). Marine seaweed cultivation is the favoured integration step in open (Petrell et al., 1993; Newkirk, 1996; Chopin et al. 1999a; b) as well as land based systems (Neori et al. , 1991; 2000; Krom et al., 1995; Vandermeulen and Gordin, 1990; Buschmann, 1996) and is widely used. An overview about the benefits of the
24 Chapter 1 application of integrated systems in aquaculture is given in Chopin et al. , 2001. Several studies have been performed to investigate the usage of macro and microalgae, hydroponics (growing plants without soil), artificial wetlands, filtering or detritivorous organisms as secondary steps in aquacultural sys tems (Schneider et al., 2005). However, the systems discussed in Schneider et al. (2005) are open systems. Thus, the research enhancing integration in aquaculture systems needs to be intensified (Costa Pierce, 2002). To date, no successful attempts have been made to set up a completely closed integrated seawater system, probably because of the lack of experience in running completely closed marine systems.
Based on the established knowledge and experience in marine recirculation systems at IFM GEOMAR the first attempt of running a completely closed integrated seawater system over a longer time period (MARE = Marine Artifi cial Recirculation System) was made (see Chapter 2 and 5) (Bischoff et al. , 2005). According to the requirements for recirculation systems all necessary treat ment steps needed to be included. The function of the swirl separator was transferred to a sedimentation tank filled with sediment. The sedimented particles could be directly used by the common ragworm Nereis diversicolor living in the sediment . Instead of a conventional biofilter a macroalgae tank was included in the first experimental phase (Wecker et al. , 2005) and re placed by a photobioreactor system for microalgae cultivation in the second experimental phase (Kube et al. , 2005; see Chapter 3 and 4) (Fig. 8). The re maining technical components in MARE were: a pump and two foam frac tionators for the removal of the suspended solids <50 m. A drawing of the system is included in Chapter 2.
25 Chapter 1
Settable Solids Dissolved Nutrients
Suspended Solids
Dissolved Nutrients
MARE System (Trail I and II) Outlook MARE System Fig. 8 Concept of the MARE system (Marine Artificial Recirculating Ecosystem): The target species Sparus aurata provides the waste (dissolved and particulate matter) being used by the other biological system components. The common ragworm ( Nereis diversicolor) is used for utilization of the particu late waste, dissolved nutrients are recycled by macroalgae (e.g. Solieria chordalis ) (Phase I) and micro algae ( Nannochloropsis spec.) (Phase II). In the present MARE system suspended solids are still re moved by foam fractionators. But future perspectives also include for example bivalves (e.g. Mytilus edulis ) for removal of particles >50 m. 1.5 Organisms Gilthead Seabream (Sparus aurata) The Gilthead seabream taxonomically can be assigned to the family Sparidae (Percoidei, Perciformes, Acanthopterygii, Teleostei, Neopterygii, Actinoptery gei, Osteichthyes). The natural habitat of Sparus aurata are sea grass beds or sandy bottoms of the coastal waters of the Mediterranean Sea and the Eastern Atlantic Ocean. Sparus aurata is a bottom dwelling species and usually lives solitary or in small shoals. This fish species reaches an average size of approx. 35 cm. In nature, Sparus aurata prefers invertebrates as food (shellfish, bivalves), but they are also considered to casually feed on plants or phytoplankton. This fish species reaches maturity after approx. one year and is a protandric hermaphrodite (changes sex from ♂♂ to ♀). In nature, spawning is restricted to the winter season but in aquacultures, perennial breeding can be observed providing controlled conditions (changes on the
26 Chapter 1 specific gravity of the water, temperature and photoperiod). Sparus aurata has planktonic larvae and the larval phase takes approx. 50 days (tempera ture 17 18° C).
The Gilthead seabream is a highly appreciated food fish, which is widely cul tivated all over the Mediterranean Sea in well organised aquacultures.
Fig. 9 Gilthead seabream ( Sparus aurata Linnaeus, 1758, left); naturally distributed in the Mediterra nean and the Eastern Atlantic Ocean and common ragworm ( Nereis diversicolor Linnaeus, 1758, right).
Common ragworm (Nereis diversicolor) The common ragworm can be classified as a member of the family Nereididae (Phyllodocida, Aciculata, Palpata, Polychaeta, Annelida).
Nereis diversicolor reaches a maximum size of approx. 10 cm and is coloured reddish brown to greenish (when reaching maturity). It is characterized by the possession of numerous paddling feet and a characteristic red line on the dorsal side. Nereis diversicolor lives along the coasts of the North Atlantic Ocean and is adjacent seas. Adaptation to brackish water and freshwater (river estuaries) has also been observed for this polychaete species. Nereis diversicolor exhibits an endobenthic way of living and mainly can be found in living tubes reaching up to 10 cm sediment depth. Food can be acquired us ing different methods. For capture of sinking particles/planktonic organ isms, a web composed of mucus is excreted into the living tube and water is pumped through this web. Thus, the mucus web acts as a sieve enriching particles/planktonic organisms. Nereis diversicolor also is able to graze on benthic algae and small benthic animals using his mandibles. Concerning
27 Chapter 1 reproduction strategy, Nereis shows a monotelic reproduction (adult worms die after reproduction).
Due to its nutrition, Nereis is especially suitable for the integration into a closed recirculation system. Particulate matter (fish faeces, food remains) can thus be removed using the natural nutrition of this worm species.
Solieria chordalis Distribution of Solieria chordalis (Solieriaceae , Rhodophyta) ranges from the Mediterranean to areas in the North Atlantic, influenced by the Gulf stream. Solieria is well adapted to temperatures demanded by the target species Gilthead seabream ( Sparus aurata ). Results about growth, reproduction and commercial cultivation were derived from an algae farm in Sylt, Germany. In contrast to other macroalgae, Solieria chordalis is well growing during the summer and resistant against biofouling.
Fig. 10 Solieria chordalis (left) (Source:www.ifremer.fr) and Nannochloropsis sp. (right) (source:www.sb roscoff.fr).
Nannochloropsis sp. Nannochloropsis (Eustigmatophyceae) is generally described as a component of the picoeukaryotic plankton because of its size range of 2 5 m (Hu and Gao, 2003). They stand at the beginning of the food chain in aquatic ecosys tems. Picoeukaryotic plankton are found throughout the world´s ocean at concentrations between 10² to 10 4 cells per cm³ in the upper photic zone (Caron et al. , 1999), playing significant roles in global carbon and mineral
28 Chapter 1 cycles (Fogg, 1995). Furthermore this microalgae contains highly nutritional compounds (e.g. sterols, polyunsaturated fatty acids; Verón et al. , 1998; Ro cha et al., 2003) and is therefore used for feeding fish larvae in mariculture. It contains only chlorophyll a (Hibbet, 1988) but also valuable pigments such as zeaxanthin, canthaxanthin and astaxanthin at high levels (Lubián et al. , 2000). 1.6 Thesis outline Based on the established knowledge and experience in marine recirculation systems at IFM GEOMAR the first attempt of running an seawater recircula tion system with different trophic levels over a longer time period (MARE = Marine Artificial Recirculation System) was made (see Chapter 2 and 5) (Bischoff et al. , 2005). The system was set up and maintained by the members of the mariculture working group. The aim of the project was to test the feasibility and the op eration ranges of such a novel recirculation system. The performance and applicability of the system should be investigated as well as the dynamics of the single modules. This work was split into three different topics (PhD the ses). The combination of these data (system and modules) are used to obtain comprehensive information about the nutrient cycling based on the daily feeding in order to evaluate a nutrient budget for the system. Hereby the fol lowing topics were in focus of the experiments: the available nutrient con centrations in each module and the extent of nutrients being used for incor poration into biomass. Further investigations dealt with the system´s limita tions and the potential dangers for the safe operation of the system. Main topic of this thesis was the development of a continuous photobioreactorsys tem based on dissolved nutrients of a marine recirculation system. The thesis is divided into five chapters (including the introduction), each fo cused on a different scientific objective.
29 Chapter 1
Major scientific objectives of this thesis were: • to investigate the applicability of a seawater recirculation system with several trophic levels (water exchange < 1% total water volume/day) for cultivation of Sparus aurata over a longer time period (chapters 2 and 5) • to determine biotic and abiotic factors limiting the performance of this system (chapters 3 and 5) • to evaluate the performance and practical applicability of the photobio reactors for Nannochloropsis sp. cultivation integrated into the system (chapter 5) • to evaluate the applicability of the foam fractionation technique for removal of bacteria and particles in a marine recirculation system suitable for microalgae cultivation (chapter 4)
30 Chapter 1 1.7 References Ackefors H. and Enell M. (1994). The release of nutrients and organic matter from aquaculture systems in Nordic countries. Journal of Applied Ichthyol ogy 10: 225 241. Barak Y. and van Rijn J. (2000). Biological phosphate removal in a prototype recirculation aquaculture treatment system. Aquacultural Engineering 22: 121 136. Benemann J.R. (1992). Microalgae aquaculture feeds. Journal of Applied Phycology 4: 233 245. Beveridge M.C.M., Philipps M.J., and Clarke R.M. (1991). A quantitative and qualitative assessment of wastes from aquatic animal production. In: Aqua culture and Water Quality, Advances in World Aquaculture, vol. 3. Edited by D.E. Brune and J.R. Tomasso. The World Aquaculture Society, Baton Rouge, LA: 506 533. Bischoff A.A. (2003). Growth and mortality of the polychaete Nereis diversi- color under experimental rearing conditions. M.Sc.thesis, Institute of Marine Research & Department of Animal Sciences, Chairgroup of Fish Culture and Fisheries, Christian Albrechts University Kiel, Germany/Wageningen Univer sity, The Netherlands; 103pp. Bischoff A.A., Kube N., Wecker B., and Waller U. (2005). MARE Marine Ar tificial Recirculating Ecosystem: Steps towards closed systems for the pro duction of marine organisms. European Aquaculture Society Special Publi cation 35: 135 136. Blancheton J.P. (2000). Developments in recirculation systems for Mediter rean fish species. Aquacultural Engineering 22: 17 31. Bock B.R. (1986). Increasing cereal yields with hihger ammonium nitrate rations – review of potentials and limitations. Journal of Environmental Sci ence And Health Part A Environmental Science and Engineering & Toxic and Hazardous Substance Control 21 (7): 723 758.
Bodvin T., Intergaard M., Norgaard E., Jensen A., and Skaar A. (1996). Clean technology in aquaculture: a production without waste products. Hydrobi ologica 327: 83 86. Bone Q., Marshall N.B., and Blaxter J.H.S. (1995). Biology of fishes. 2 nd edi tion. Blackie London, New York, 332 pp. Bothe H, Jost G and Schloter M. (2000). Molecular analysis of ammonia oxi dation and denitrification in natural environments. FEMS Microbiology Re views 24 (5): 673 690.
31 Chapter 1
Bovendeur J., Eding E.H., and Henken A.M. (1987). Design and performance of a water recirculation system for high density culture of the African Cat fish, Claria gariepinus (Burchell 1822). Aquaculture 63: 329 – 353. Bower C.E. (1978). Ionization of ammonia in seawater: Effect of temperature, pH, and salinity. J.Fish.Res.Board Can. 35:1012 1016. Boyd C.E. (2003). Guidelines for aquaculture effluent management at the farm level. Aquaculture 226: 101 112. Braaten B., Aure J., Ervik A., and Boge E. (1983). Pollution problems in Norwegian fish farming. ICES CM 1983/F 26: 1 11. Braaten B., Poppe T., Jacobsen P., and Maroni K. (1986). Risks from self pollution in aquaculture: evaluation and concequences. In: Efficiency in Aquaculture Production: Disease and Control. Edited by E. Grimaldi and H. Rosenthal. Proceedings of the 3 rd International Conference on Aquafarming ‘Aquacoltura ´86’, Verona, Italy, Oct. 9 10, 1986: 139 165. Braaten B. (1992). Impact of pollution from aquaculture in six Nordic coun tries. Release of nutrients, effects and waste water treatment. In: Aquacul ture and the Environment. Edited by N. DePauw and J. Joyce. EAS Special Publications 1992, No. 16, Gent, Belgium: 79 101. Brown M.R., Jeffrey S.W., and Garland C.D. (1989). Nutritional Aspects of Microalgae used in Mariculture; a Literature Review. CSIRO Marine Labora tories Report 205, 44pp. Buschmann A.H., Troell M., Kautsky N., and Kautsky L. (1996). Integrated tank cultivation of salmonids and Gracilaria chilensis (Rhodophyta). Hydro biologica 326/327: 75 82. Caron D.A., Peele E.R., Lim E.L., and Dennet M.R. (1999). Picoplankton and nanoplankton and their trophic coupling in the surface waters of the Sar gasso Sea south of Bermuda. Limnol. Oceanogr. 44(2): 259 272. Chamberlain G. and Rosenthal H. (1995). Aquaculture in the next century: Opportunities for growth challenges of sustainability. World Aquaculture 26: 21 25. Chopin T., Buschman A.H., Halling C., Troell M., Kautsky N., Neori A., Kraemer G.P., Zertuche Gonzales J.A., Yarish C., and Neefus C. (2001). Inte grating seaweeds into marine aquaculture systems: A key toward sustain ability. Journal of Phycology 37: 975 986. Chopin T., Sharp G., Belyea E., Semple R., and Jones D. (1999a). Open wa ter aquaculture of the red algae Chondrus crispus in Prince Edward Island, Canada. Hydrobiologica 398/399: 417 425.
32 Chapter 1
Chopin T., Yarish C., Wilkes R., Belyea E., Lu S., and Mathieson A. (1999b). Developing Porphyra /salmon integrated aquaculture for bioremediation and diversification of the aquaculture industry. Journal Applied Phycology 11: 463 472. Colt J. and Armstrong D. (1979). Nitrogen toxicity to fish, crustaceans and molluscs. Department of Civil Engineering, University of California, Davis, California. Cripps S.J. and Bergheim A. (2000). Solids management and removal for in tensive land based aquaculture production systems. Aquacultural Enginee ring 22: 33 56. Cross D. (2003). Statistik kurz gefasst – Landwirtschaft und Fischerei 34/2003. Statistisches Bundesamt. EDS – Europäischer Datenservice www.eds destatis.de Davenport J., Black K., Burnell G., Cross T., Culloty S., Ekaratne S., Fur ness B., Mulcahy M., and Thetmeyer H. (eds) (2003). Aquaculture – the eco logical issues. The British Ecological Society – Ecological Issues Series. Blackwell Publishing 89pp. EIFAC (1986). Flow through and Recirculation Systems. Report of the work ing group on terminology, format and units of measurement. EIFAC Techni cal Paper 49. FAO (2002, 2005) www.fao.org. FAO, 1988. Aspects of FAO´s policies, programmes, budget and activities aimed at contributing to sustainable development. Document to the ninety fourth Session of the FAO Council, Rome, 15 25 November 1988. Rome, FAO. CL 94/6. Fogg G.E. (1995). Some comments on picoplankton and its importance in the pelagic ecosystem. Aquat. Microb. Ecol. 9 (1): 33 39. GESAMP (IMO/FAO/UNESCO IOC/WMO/WHO/IAEA/UN/UNEP Joint Group of Experts on the Scientific Aspects of Marine Environmental Protec tion), 2001. Planning and management for sustainable coastal aquaculture development. Rep.Stud.GESAMP, 68: 90pp. Gladue R.M. and Maxey J.E. (1994). Microalgal feeds for aquaculture. Jour nal of Applied Phycology 6: 131 141. Hagopian D.S. and Riley J.G. (1998). A closer look at the bacteriology of ni trification. Aquacultural Engineering 18: 223 244. Hall P.O.J., Kollberg S., and Samuelsson M.O. (1992). Chemical fluxes and mass balance in a marine fish cage farm. IV.Nitrogen. Marine Ecology Pro gress Series 89: 81 91.
33 Chapter 1
Hargreaves J.A. (1998). Nitrogen biogeochemistry of aquaculture ponds. Aquaculture 166: 181 212. Hibberd D.J. (1988). Eustigmatophyceae. In: Handbook of Protoctista. Mar gulis L., Corliss J., Melkonian M. and Chapman D. (eds). Jones & Bartlett Publ., Boston, 30 50. Hussenot J., Lefebvre S., and Brossard N. (1998). Open air treatment of wastewater from land based marine fish farms in extensive and intensive systems: current technology and future perspectives. Aquatic Living Re sources 11 (4): 297 304. Hu H. and Gao, K. (2003). Optimization of growth and fatty acid composition of a unicellular marine picoplankton, Nannochloropsis sp., with enriched carbon sources. Biotechn. Lett., 25(5): 421 425. Kioussis D.R., Wheaton F.W., and Kofinas P. (2000). Reactive nitrogen and phosphorus removal from aquaculture wastewater effluents using polymer hydrogels. Aquacultural Engineering 23: 315 332. Krom M.D., Porter C., and Gordin H. (1985). Nutrient budget of a marine fish pond in Eilat, Israel. Aquaculture 51: 65 80. Krom M.D. and Neori A. (1989). A total nutrient budget for an experimental intensive fishpond with circularly moving seawater. Aquaculture 83: 345 358. Kube N., Bischoff A.A., Wecker B., and Waller U. (2005). The feasibility of a photobioreactor (microalgae) for removal of dissolved nutrients in a closed recirculating system. European Aquaculture Society Special Publication 35: 289 290. Liao P.B. and Mayo R.D. (1974). Intensified fish culture combining water re conditioning with pollution abatement. Aquaculture 3: 61 85. Liltved H. and Cripps S.J. (1999). Removal of particle associated bacteria by prefiltration and ultraviolet irradiation. Aquacult. Res. 30: 445 450. Losordo T.M., Masser M.P., and Rakocy J. (1998). Recirculation Aquaculture Tank Production systems: an overview about critical considerations. SRAC Publication 451. Losordo T.M., Masser M.P., and Rakocy J.E. (1999). Recirculating aquacul ture tank production systems: A review of component options. SRAC Publication 453. Lubián L.M., Montero O., Moreno Garrido I., Huertas E., Sobrino C., Gonzalés Delvalle M., and Parés G. (2000). Nannochloropsis (Eustigmato phyceae) as a source of commercially valuable pigments. J. Applied Phycol., 12 (3) (5): 249 255.
34 Chapter 1
Lupatsch I. and Kissil G.W. (1998). Predicting aquaculture waste from Gilt head Seabream ( Sparus aurata ) culture using a nutritional approach. Aquat. Living Ressour. 11: 265 268. Masser M.P., Rakocy J., and Losordo T.M. (1999). Recirculation Aquaculture Tank Production Systems: Management of recirculation systems. SRAC Pub lication 452. Mourente G., Lubián L.M., and Odriozola J.M. (1990). Total fatty acid com position as a taxonomic index of some marine microalgae used as food in marine aquaculture. Hydrobiologica 203: 147 154. Nash C.E., Burbridge P.R., and Volkman J.K. (eds) (2005). Guidelines for Ecological Risk Assessment of Marine Fish Aquaculture. NOAA Technical Memorandum NMFS NWFSC 71. Prepared at the NOAA Fisheries Service Manchester Research Station International Workshop April 11 14, 90pp. Neori A., Cohen I., and Gordin H. (1991). Ulva lactuca biofilters for marine fish pond effluents. II. Growth rate, yield and C:N ratio. Botanica Marina 34: 483 489. Neori A., Shpigel M., and Ben Ezra, D. (2000). A sustainable integrated sys tem for culture of fish, seaweed and abalone. Aquaculture 186: 279 291. Newkirk G. (1996). Sustainable coastal production systems: a model for in tegrating aquaculture and fisheries under community management. Ocean Coast. Manag. 32: 69 83. Otte G. and Rosenthal H. (1979). Management of a closed brackish water system for high density fish culture by biological and chemical water treat ment. Aquaculture 18: 169 181. Petrell R.J. and Alie S.Y. (1996). Integrated aquaculture of salmonids and seaweeds in open systems. Hydrobiologica 326/327: 67 73. Pfeiffer T.J. and Rusch K.A. (2000). An integrated system for microalgal and nursery seed clam culture. Aquacultural Engineering 24: 15 31. Piedrahita R.H. (2003). Reducing the potential environmental impact of tank aquaculture effluents through intensification and recirculation. Aquaculture 226: 35 44. Pinto C.S.C., Souza Santos L., and Santos P.J.P. (2001). Development and population dynamics of Tisbe biminiensis (Copepoda: Harpacticoida) reared on different diets. Aquaculture 198: 253 267. Renaud S.M., Luong Van Thinh, Parry D.L. (1999). The gross chemical com position and fatty acid composition of 18 species of tropical Australian mi croalgae for possible use in mariculture. Aquaculture 170: 147 159.
35 Chapter 1
Renaud S.M., Parry D.L., and Luong Van Thinh (1994). Microalgae for use in tropical aquaculture I: Gross chemical and fatty acid composition of twelve species of microalgae from the Northern Territory, Australia. Journal of Ap plied Phycology 6: 337 345. Rheinheimer G., Hegemann W., Raff J., and Sekoulov I. (1988). Stickstoff kreislauf im Wasser: Stickstoffumsetzungen in natürlichen Gewässern, in der Abwasserreinigung und Wasserversorgung. R. Oldenburg Verlag GmbH, München. Rocha J.M.S., Garcia J.E.C., and Henriques M.H.F. (2003). Growth aspects of the marine microalgae Nannochloropsis gaditana . Biomolecular Engineer ing 20: 237 242. Rosenthal H. (1997). Aquaculture and sustainable development. In: Summit of the sea, Understanding and managing the oceans. CORE: 92 96. Rosenthal H. and Grimaldi E. (1990). Efficiency in Aquaculture production: Technology improvements in farming systems. Proc. 4th Intern. Conf. Aqua farming "AQUACULTURA 88". Verona: Ente Fiere di Verona. Rosenthal H., Hoffmann R., Jörgensen L., Krüner G., Peters G., Schlotfeldt H. J., and Schomann H. (1982). Water management in circular tanks of a commercial intensive culture unit and its effects on water quality and fish condition. ICES Statutory meeting, C.M. 1982/F:22, 13pp. Russo R.C. and Thurston R.V. (1991). Toxicity of ammonia, nitrite and ni trate to fishes. In: Aquaculture and Water Quality. Edited by E.D.Brune and J.R. Tomasso. The World Aquaculture Society, LA: 58 89. Schlegel H.G. (1992). Allgemeine Mikrobiologie. 7.Auflage. Thieme Stuttgart New York. Schmidt I., Gries T., and Willuweit T. (1999). Nitrification Fundamentals of the metabolism and problems at the use of ammonia oxidizers. Acta Hydro chimica et Hydrobiologica 27 (3): 121 135. Schneider O., Sereti V., Eding E.H., and Verreth J.A.J. (2005). Analysis of nutrient flows in integrated intensive aquaculture systems. Aquacultural Engineering 32: 379 401. Støttrup J.G. and McEvoy L.A. (eds.) (2002). Live feeds in marine aquacul ture. Blackwell publishing, 336pp. Støttrup J.G. and Norsker N.H. (1997). Production and use of copepods in marine fish larviculture. Aquaculture 155: 231 247. Strous M., Fuerst J.A., Kramer E.H.M., Logemann S., Muyzer G., Van de Pas Schoonen K.T., Webb R., Kuenen J.G.,and Jetten M.S.M. (1999). "Miss ing lithotroph identified as new plantomycete." Nature 400: 446 449.
36 Chapter 1
Summerfelt S.T. (2002). An integrated approach to aquaculture waste man agement in flowing water systems. Proceedings of the 2nd International Con ference on Recirculating Aquaculture: 87 97. Tiedje J. M. (1988). Ecology of denitrification and dissimilatory nitrate reduc tion to ammonium. Wageningen, NL. Troell M., Halling C., Neori A., Chopin T., Buschman A.H., Kautsky N., and Yarish C. (2003). Integrated mariculture: asking the right questions. Aqua culture 226: 69 90. Trussel, R.P. (1972). The percent un ionized ammonia in aqueous ammonia solutions at different pH levels and temperatures. J.Fish. Res. Board Can. 29: 1505 1507. Vandermeulen H., and Gordin H. (1990). Ammonium uptake using Ulva (Chlorophyta) in intensive fishpond systems: mass culture and treatment of effluent. Journal Applied Phycology 2: 363 374. Waller U. (2000). Tank culture – including raceways and re circulating sys tems. In: Environmental impacts of aquaculture. Edited by K.D.Black. Shef field Academic Press. Waller U., Sander M., and Piker L. (2001). Low energy and low water con sumption recirculation system for marine fish: first results from a test run with Dicentrarchus labrax in an improved recirculating system and sugges tions on an integration into secondary production lines. European Aquacul ture Society Special Publications 29: 265 266. Waller U., Bischoff A.A., Orellana J., Sander M., and Wecker B. (2003). An advanced technology for clear water aquaculture recirculation systems: Re sults from a pilot production of Sea bass and hints towards "Zero Dis charge". European Aquaculture Society Special Publications 33: 356 357. Waller U., Sander M., and Orellana J. (2005). A “low energy” commercial scale recirculation system for marine finfish. European Aquaculture Society Special Publications 35: 459 460. Wecker B. (2002). Anorganische Stoffflüsse in einer experimentellen Kreis laufanlage mit definiertem Fischbesatz. Diplomarbeit, Institut für Meeres kunde Kiel, Germany. Wecker B., Bischoff A.A., He J., Kube N., Lüning K., and Waller U. (2005). Modelling the nutrient uptake and benefits of seaweed filter integrated in a closed recircualting system. European Aquaculture Society Special Publica tion 35: 465 466. Welch E.B. and Lindell T. (1992). Ecological Effects of Wastewater. Applied Limnology and Pollutant Effects. Chapman and Hall, London: 76 81.
37 Chapter 1
Weston D.P. (1991). The effects of aquaculture on indigenous biota. In: Aquaculture and Water Quality. Edited by E.D. Burne. World Aquaculture Society, Batan Rouge, LA: 534 567. Wickens J.F. (1980). Water quality requirements for intensive aquaculture: a review. EIFAC 80:1 17. Wrage N. and Velthof G.L. (2001). "Role of nitrifier denitrification in the pro duction of nitrous oxide." Soil Biology & Biochemistry 33: 1723 1732. Zumft W. G. (1997). "Cell biology and molecular basis of denitrification." Mi crobiology and Molecular Biology Reviews 61(4): 533 616.
38
Chapter 2 MARE – Marine Artificial Recirculated Ecosystem: feasibility and modelling of a novel integrated re circulation system Wecker B., Kube N., Bischoff A.A., and Waller U. (2006)
39 Chapter 2 MARE – Marine Artificial Recirculated Ecosys tem: feasibility and modelling of a novel inte grated recirculation system Wecker, B., Kube N., Bischoff A.A., Waller U. Abstract In conventional recirculating systems most of the nutrients supplied with the food are discharged because such installations do not include treatment steps for nutrient recycling. An advanced farming design for low water dis charge (<< 1 % d 1 system volume) was investigated. The MARE system is based on the concept of a closed biological integrated recirculating system not only with low water discharge but also with nutrient recycling. The target organism was Gilthead seabream ( Sparus aurata ). In secondary production compartments of the MARE system particulate matter was used by detritivo rous worms ( Nereis diversicolor ) and dissolved nutrients were used by macroalgae ( Solieria chordalis ). There was no additional technical biofilter system. This paper describes the feasibility of the MARE system in order to maintain the water quality within safe limits and support adequate fish growth. Sec ondly, a mathematical model was developed to understand the nutrient flows and biomass developments within the MARE System. It was possible to es tablish a nutrient budget for MARE. The results show that nitrogen is mostly converted microbiologically by nitrification and denitrification. Phosphorus was mainly taken up by macroalgae. The scientific concept for the MARE system was shown to be feasible: a simplified recirculated ecosystem can be used to recycle nutrients and to maintain water quality. However, the secon dary modules ( Nereis, Solieria ) were found to be less important for the nutri ent retention, but increased valorisation of the system. Thus, integration of different trophic levels led to an increased nutrient utilization.
Keywords: marine recirculation system, integration, nutrient budget model, Nereis diversicolor , Sparus aurata , Solieria chordalis , artificial ecosystem
40 Chapter 2 2.1 Introduction In closed recirculation systems waste is usually concentrated and dis charged (solids in sinks) or dissolved nutrients are accumulating (nitrate and phosphate). The daily amount of waste is rather high: results from different aquacultural production systems showed that only 20 30% of nitrogen in the feed are used by fish to synthesize biomass (Krom et al. , 1985; Krom and Neori, 1989; Hall et al., 1992; Lupatsch and Kissil, 1998; Hargreaves, 1998). Utilization of phosphorus is between 10 30% of feed intake (Krom et al., 1985; Krom and Neori, 1989; Barak and van Rijn, 2000; Lupatsch and Kis sil, 1998). Thus, 70 90% of the phosphorus provided with the feed are ex creted either in dissolved or particulate form. To date, the operation of recirculation systems is focused on recycling the water only. Nutrients like nitrogen, phosphorus or carbon compounds are eliminated by water treatment steps either biologically (biofiltration) or by water exchange, (Losordo et al. , 1999; Waller et al. , 2005). Thus, these valu able organic compounds are lost for the system and released having enor mous impacts on the environment (e.g. eutrophication). In modern aquacul ture systems a comprehensive nutrient recycling should be maintained in order to reduce these environmental impacts. Waste from fish production should be understood as „new resources“ (Chamberlain and Rosenthal, 1995). Due to the better utilization of these resources by integration of dif ferent trophic levels, the profitability of recirculation systems can be en hanced (Asgard et al. , 1999; Schneider et al. , 2005). Additional biomass supports the economical diversification and the benefit per production unit (Chopin et al. , 2001). Production systems without any nutrient and energy recycling are supposed to have less chances in future (Troell et al. 2003). However, integration is not a new idea because in freshwater and brackish aquaculture systems integrated aquaculture has been practiced for centuries (Chopin et al. , 2001). Marine seaweed production is the mainly used integra tion step in open marine (Petrell et al., 1996; Newkirk, 1996; Chopin et al. 1999a; b) as well in land based systems (Neori et al. , 1991; 2000; Krom et
41 Chapter 2 al., 1995; Vandermeulen and Gordin, 1990; Buschmann, 1996). An overview about the current discussion concerning the integration of macroalgae into these systems is given in Chopin et al. (2001). Several studies describe the usage of macro and microalgae, hydroponics (growing plants without soil), artificial wetlands, filtering or detritivorous organisms as secondary steps (Schneider et al., 2005 and references therein), but these systems are char acterized by larger water exchanging rates.
Based on the knowledge of low discharge (<1% per day) recirculation sys tems, the goal of this study was to develop a completely new recirculation system being characterized by a closed water recirculation as well as by nu trient recycling via the integration of secondary steps ( Nereis, Soleria ).
2.2 Material and Methods 2.2.1 MARE-System MARE (4.5m³) consisted of different units (Fig. 1): two fishtanks (700L vol ume) (1) each stocked with 85 juvenile Gilthead seabream ( Sparus aurata ) with an average weight of 66.5 ± 11g. A circular current was established by airlifts. Gilthead Seabream was chosen due to its comparatively easy cultiva tion and its wide range of tolerated salt concentrations and temperatures. Water from the fish tanks was completely transferred to the detritivorous re actor (2). The bioreactor for detritivorous organisms (2.1m² sediment surface, water column 0.7m) was filled with a 0.1m deep sand layer (grain size ≤ 2mm) and stocked with Nereis diversicolor at a density of approx. 900 950 individuals per m² on 18.6.04. Additional walls were installed within the tank to elongated the water route through the tank and thus to enhance the settling of suspended solids. For the removal of dissolved nutrients a macro algae cultivation unit (2.3 m² surface area) was included with a stocking density of 4.4 kg/m² surface (3) of floating Solieria chordalis in an aerated circular tank with 400 E m 2 s 1 illumination (day:night 16:8 hours). A pump (type AG8, ITT Hydroair international, Denmark) (4) provided two foam fractionators (type Outside Skimmer III; Erwin Sander Elektroapparate
42 Chapter 2
GmbH; Uetze Eltze, Germany) with an average flow rate of 1000L*h 1 each (6) as well as the two fishtanks with about 600 to 800 L*h 1 respectively. The protein skimmers were aerated with compressed air and an additional ozone addition, produced by an ozone generator (Sander, Ozonizer A2000). In a separate bypass an additional tank filled with 50 litres of freshwater was at tached to the foam fractionators for rinsing the foam collectors automatically every 15 minutes. Removed foam was collected in the freshwater tank. When the MARE experiment started, Gilthead seabream were already cul tured for 182 days within the MARE system. Also Nereis diversicolor was stocked one month before the start of the experiments at 11.11.04. At the beginning, a biofilter for nitrification was included for safety reasons, but it was shut down during the experimental period (11.11.2004 – 01.06.2005).
43 Chapter 2
1) 2) 3) 5)
400
1000 A Z
1670 Z 1700 0
0 Z 3 1 0
3 A Z 0 7 A 0 1 0 0 0 1 9 0 1 0 1 1 A
Fig. 1 Flow chart of the MARE system (Marine Artificial Recirculating Ecosystem) and 3D figures of the modules. All units of the module are given in mm. (1) = fishtanks ( Sparus aurata ), (2) = detritivo rous culture tank ( Nereis diversicolor ) filled with sediment and additional wall segments; (3) = macro algae tank ( Solieria chordalis ) with a central aeration; (4) = pump; (5) = foam fractionators, arrows indicate water flow, double triangle = valve, Z = inlet, A = outlet
2.2.2 Measurements and Methods Biomass determination Fish were fed with pellets of 4.5mm size (Biomar, Aqualife 17, see Table 1). The daily food rate was adjusted following the results of fish biomass deter mination (wet weight of at least 30 individuals per tank) at intervals of appr. 3 weeks.
44 Chapter 2
The Nereis biomass was also determined (number and wet weight of worms) using four subsamples. Therefore, the detritivorous tank was divided into four parts. From each part, a core of approx. 800 cm 3 was sampled. To re move the sediment and to enrich worms, a sieve of 1 mm mesh size was used for sampling.
Wet weight of Solieria was determined at an average of 1 to 2 weeks. Yield of biomass was remained in the tank until reaching stocking density of 8kg/m² at week 20 of the experimental period. Afterwards the stocking density was remained constant at this value by removing the weekly gain from the sys tem.
Water parameters Water samples for analysis of dissolved nutrients were taken daily at three different points of the MARE system: outlet fishtanks, outlet Nereis tank and outlet algae tank. Water samples were stored at –20°C and analysed with an AA3 Autoanalyzer (Method no. G 016 91 for Nitrate N, G 029 92 for Nitrite N, G 102 93 for Ammonia N, G 103 93 for Phosphate, Bran Lübbe GmbH; Norderstedt, Germany). Flow rates through fish tanks and foam fractionators were recorded and ad justed to 600L h 1 and 1000L h 1, respectively. Online measurements of re dox potential, pH and dissolved oxygen were recorded with a control module KM 2000 (Meinsberg) and a portable measuring device (WTW multi 350). System water level was controlled daily and if necessary, adjusted with fil tered brackish seawater. Salinity of the system water was 24.1 ± 0.8 psu. If needed, artificial sea salt was used to increase salinity. The foam collectors were cleaned daily and the freshwater for rinsing was changed. The daily loss of water via the foam fractionators was recorded by the increased water volume in the freshwater tank.
45 Chapter 2
Solid components The analysis of solid waste, tissues and sediment parameters were per formed as follows: dry matter content was determined by drying the sample in a drying oven at 60°C overnight. Organic content was measured by incin eration of organic matter in a muffle furnace. C/N ratio was determined by an element analyzer (type NA 1500 series 2, FISONS). Calorific value was measured by complete sample combustion using an IKA calorimeter C4000. Weighing was performed using with a Sartorius A 210 P (max. 200g) and a Sartorius U 4600 P (max. 4000g). Every week a sample of the rinsing water of the foam fractionators was taken to determine the proportion of suspended solids removed from the system. 12 tubes of 10ml sample were centrifuged and the supernatant was stored for later water analysis. Dry matter and C/N ratio analysis of the pellets re sulting from centrifugation were performed analogous to the analysis of solid components. The organic matter content of the sediment also was analysed weekly with 5 subsamples of appr. 10 cm³ sediment according to this princi ple 2.2.3 Modelling The aim of this study was not only to test the feasibility of the MARE concept but also to process the experimental data within a mathematical model in order to evaluate the nutrient flows and biomass developments within the MARE system. Mathematical models are integrative and important tools of interdisciplinary work in order to combine knowledge and scientific methods from different subjects or data from diverse sources. A model can simplify complex processes for better understanding and helps to concentrate on the essential processes. In summary, developing a model demands comprehen sion but also augments knowledge (Ebenhöh, 2004).
46 Chapter 2
Fig. 2 shows the concept of the numeric model developed for MARE. Vari ables and parameters are listed in Tab. 1 and indicate the connections among the respective modules.
Fig. 2 Flow chart of the nutrient fluxes in the MARE system. The concept of the numeric model and the connections of the single modules are shown schematically. The marked boxes indicate the final values of the nutrient budget. The model was designed to analyse nitrogen and phosphorus as nutri ents of major importance for the system.
47 Chapter 2
Tab. 1 Denomination of variables and parameters, dimensions, description and source used for numeric MARE model, t denotes production time in days, x denotes kind of nutrient Denomination Dimension Description Source Wfish,t g individual weight of fish experiment t denotes time in days (d) fish,t d 1 specific growth rate calculated Yfish,t g d 1 yield, weight increment calculated F,t g d 1 weight of given aquafeed per day calculated FCR t feed conversation ratio calculated intake nutrients g d 1 total amount of nutrients introduced to recirculation calculated system per day Cx, feed % percentage of nutrient in given feed literature x denotes the kind of nutrient Faeces x,t g amount of nutrients in faeces literature ADC x % apparent digestibility coefficient literature faeces soluble x,t g dissolved fraction of faeces due to leaching literature faeces unsolu g unsoluble fraction of faeces after leaching literature ble x,t RM mean,x % remaining matter of faeces after leaching literature excretion x,t g amount of dissolved nutrient from fish metabolism calculated retention x,t g amount of nutrient remained in body tissues calculated suspended sol g amount of nutrient in particles separated by foam frac experiment ids x,t tionation efficiency foam % share solids removed by foam fractionation process calculated fractionator SettableSolids x,t % share of solids removed by swirl separation calculated worm,t d 1 growth rate of polychaetes depending on the energy literature supplied with fish faeces EsettableSolids,t kJ d 1 daily amount of energy available from settable solids for literature worm 1 each individual worm CE,settableSolids kJ Energy content of settable solids experiment Yworm,t g yield, weight increment of worms in the detritivorous calculated reactor Wworm,t g individual weight of worms experiment nworm,t number of worms in the detritivorous reactor experiment MN,worm,t nworm,t d 1 natural worm mortality calculated EnergyloadTotal t KJ d 1 total energy available for production in the detritivorous reactor derived from faeces and worm mortality cDM, worm g kg 1 dry matter of worm literature Eworm KJ g 1 DW energy content of worm literature MC,worm,t nworm, t d 1 worm mortality due to cannibalistic behaviour calculated cTAN,t mg L 1 TAN concentration in the recirculation water calculated rTAN,algae,t g TAN h 1 daily TAN uptake by macroalgae experiment m 2 rmax,TAN,algae g TAN h 1 maximum daily TAN uptake by macroalgae experiment m 2 rTAN%,algae,t % TAN h 1 relative daily TAN uptake by macroalgae experiment m 2 rmax,TAN%,algae % TAN h 1 maximum relative daily TAN uptake by macroalgae experiment m 2 BTAN,t g m 2 h 1 amount of ammonia available in the macroalgae reactor calculated yalgae,t g yield, biomass increase in the macroalgae reactor calculated rP,algae,t g P h 1 m 2 phosphorus uptake of algae calculated rnitrification,t mg h 1 m 2 microbial TAN oxidation calculated BOM,t g organic matter available for denitrification calculated rdenitrification,t mg h 1 m 2 microbial removal of nitrate calculated z ratio of required organic matter to reduce nitrate to ni calculated trogen gas ct mg L 1 nitrate concentration in the system water calculated rN,bacteria,t mg NO 3 h 1 nitrate assimilation by bacteria calculated m 2 rP,bacteria,t mg PO 4 h 1 phosphorous assimilation by bacteria calculated m 2 cP,t mg L 1 phosphate concentration in the system water calculated
48 Chapter 2
Module Fish a) Fish growth and feeding rate Firstly, fish assimilation results in an increase of fish biomass. In the model, the specific growth rate per day ( fish, t ) was determined by W = fish ,t+1 µ fish ,t Ln (Equ. 1) W fish ,t where W fish,t = wet weight fish, t day of experiment
fish,t was derived from experimental data of wet fish weight (see section 3.1 fish growth ).
The yield of fish (Y fish,t ) per day can be described by = × Y fish ,t µ fish ,t W fish ,t (Equ. 2)
The relationship between feeding rate (F t) and specific growth rate ( fish,t ) can be described as feed conversion ratio (FCR t)
= Ft FCR t (Equ. 3) Yt,Fisch and hence, daily feeding rate (F t) is calculated by = × Ft FCR t Yt, fish (Equ. 4)
b) Fish faeces Fish do not use the complete amount of N and P of the provided feed for syn thesis of biomass. Over 70% are excreted as dissolved nutrients (esp. ammo nia nitrogen) and solids. These unused nutrients form the basis for the inte grated modules “detritivorous culture tank” and “macroalgae”.
49 Chapter 2
The total nutrient entry (entry nutrients ) into the recirculation system can be calculated from the feeding rate per day (F t) and the nutrient concentration of the feed (c x,feed ): F × c entry = t x, feed (Equ. 5) nutrients ,x 100 where c x is the fraction of each nutrient component (protein, PON, POP, or ganic matter (OM) or dry weight (DW) of the feed, Tab. 1). Tab. 1 Feed composition (BIOMAR Ecostart 17), apparent digestibility coefficients (ADC, 20 24°C) and nutrient concentrations of Seabream ( Sparus aurata ) (Lupatsch & Kissil, 1998) fraction in feed Digestibility fraction in fish nutrient [x] [c x,feed in %] [ADC x in %] [c x,fish in %] protein 50 82.8 17.8 nitrogen (PON) 8 82.8 2.85 Phosphor (POP) 1.4 47.8 0.72 org. matter (OM) 79.9 68.9 dry weight (DW) 90.4 60.7 Faeces are defined as undigested remains of the feed. Based on experimen tally derived apparent digestibility coefficients (ADC) for Gilthead seabream (Lupatsch & Kissil, 1998) and the composition of the feed (Tab. 1) the amount of expected solid faeces per day (faeces x,t ) for each nutrient was cal culated by
100 − ADC faeces = entry × x (Equ. 6) x,t nutrients ,x 100 Once released into water, the excreted faeces are subject to gradual chang ing of nutrient composition with time due to passive leaching. Leaching processes are considered to be terminated after 6 hours retention time in water (Lupatsch & Kissil, 1998; Tab. 2). For the numeric model average numbers for the remaining matter (RM mean ) were used.
50 Chapter 2
Tab. 2 Remaining matter (RM, %) of the solid faeces for each nutrient (x) after 6h, 24h and 48h water retention time for Gilthead seabream (following Lupatsch & Kissil, 1998) RM after 6 RM after 24 RM after 48 nutrient [x] RM mean hours hours hours nitrogen (PON) 57.4 54.9 56.8 56.4 phosphor (POP) 88.3 81.5 84.0 84.6 org. matter 59.0 57.8 57.8 58.2 (OM) dry weight (DW) 65.0 63.7 64.6 64.4
Due to leaching, faeces can be divided into a soluble part (faeces soluble ) and unsoluble part (faeces unsoluble ) by faeces ×(100 − RM ) faecesSo lub le = t,x mean (Equ. 7) x,t 100 faeces × RM faecesUnso lub le = t,x mean , respectively (Equ. 8) x,t 100 c) fish biomass synthesis (= retention) Retention of N and P in fish biomass was assessed by considering the yield of fish (Y t,fish ) and the nutrient fraction in fish (c x,fish ; Tab. 1) by × Y fish ,t c x, fish retention = (Equ. 9) x,t 100 d) excretion In contrast to faeces, all metabolic products from fish are termed as excre tion. The excreted dissolved proportions of N and P in form of total ammonia nitrogen (TAN) and orthophosphate were determined by the following equa tion: × Ft c x, feed excretion = − retention − faeces (Equ. 10) x,t 100 x,t x,t
51 Chapter 2
For the model, a minor importance of excreted urea or the rapid conversion of urea to ammonia was assumed. Consequently, total ammonia nitrogen (TAN) is considered to be the only excreted dissolved nitrogen compound.
Module Foam Fractionation Total daily amount of faeces (dry weight, DW) in relation to total nutrient en try (DW) (Equ. 5) is 39.3% calculated according to Equ. 6 (faeces DW,t ) and Tab. 1 (digestibility of DW). The unsoluble fraction can be determined ac cording to Equ. 8 (faeces unsoluble DW,t ) and Tab. 2 (RM mean DW) 25,3%. The amount of unsoluble faeces can be divided into a fraction of settable solids and a fraction of suspended solids. The fraction of suspended solids is removed from the process water by foam fractionation. This process is considered to be very effective, because the turbidity in a closed recirculation system with a foam fractionator for re moval of suspended solids is at constant low levels of <7mg dry weight L 1 (Waller et al., 2003). The efficiency of the foam fractionators was used to determine the amount of suspended solids as fraction of nutrient entry (entry nutrients ). Assuming the amount of suspended solids separated by foam fractionation to correspond to the daily load of suspended solids in the recirculation system (experimen tal observations, this study), the following equation can be used:
cx, feed Efficiency FoamFracti onators SuspendedS olids = F × × (Equ. 11) x,t t 100 100 Data concerning dry weight and C/N ratio of removed suspended solids were available from rinsing water samples of the foam fractionators: regression analysis of experimental data (removed suspended solids vs. nutrient up take) resulted in 9.5% for the efficiency of foam fractionation (see Fig. 9, subsection 3.2.1 Module foam fractionator ).
52 Chapter 2
Module Detritivorous Reactor The proportion of settable solids can be estimated according to Equ. 11 as suming 100% of settable solids to be found in the detritivorous reactor:
c 25 3, − Efficiency FoamFracti onators SettableSo lids = F × x,Futter × (Equ. 12) x,t t 100 100 This proportion is used as nutrition source by Nereis diversicolor . The rela tive growth rate of the polychaete is basically depending on the amount, the nutrient composition and the energy content of the fish faeces (Bischoff, 2003). Growth experiments with faeces from a swirl separator of a closed re circulation system showed that growth rates up to 2,8% d 1 are possible, if the feed energy content is sufficient (Bischoff, 2003).
Growth of worms ( worm,t ) can be described by (µ − µ )× E µ = max min settablesS olids ,t + µ (Equ. 13) worm ,t ± min K s EsettableSo lids ,t where E settableSolids,t is the daily energy load derived from settable solids per individual worm [KJ d 1 worm 1], max the maximum growth rate [at E settableSol ids,t = ∞ in % d 1], min the minimum growth rate [at E settableSolids,t = 0 in % d 1] and K S is the Michaelis constant [KJ d 1 worm 1].
The energy load of settable solids (DW) (E settableSolids DM,t ) reaching the detri tivorous reactor is defined by: SettableSo lids × c = DM ,t E,settableSo lids EsettableSo lids ,t (Equ. 14) nworm ,t where c E is the energy content of the settable solids (C E,settableSolids ) and the number of worms (n worm,t) in the detritivorous tank.
53 Chapter 2
The yield (weight increment) of worm (Y worm,t ) can be calculated by using Equ. 13 and 14: µ ×(W ⋅n , t) Y = worm ,t worm ,t worm (Equ. 15) worm ,t 100
The average wet worm weight (W worm,t ) at any production day (t) is:
W × µ W × µ worm ,t−1 worm ,t + worm ,t worm ,t 100 100 W = W − + (Equ. 16) worm ,t worm ,t 1 2 Equ. 16 contains a mathematical operation, if is negative: the daily yield is summed with the absolute value and afterwards divided by 2 to ensure that the average worm weight can not decrease. Thus, in the model the smallest Y is 0. This approximation was performed because it can be assumed that Ne- reis sp. does not build up considerable biological reserves (e.g. lipids) to en dure starvation periods and therefore die after rather short starvation peri ods. Accordingly, the model converts negative into a natural mortality rate per day (M N,worm,t ):
Y − Y = worm ,t worm ,t × 1 M N ,worm ,t (Equ. 17) 2 Wworm ,t The same mathematical operation ensures, that only <0 are converted to dead individuals per day. In this case n worm,t is decreasing calculated by: = − nWorm ,t nworm ,t−1 M N ,t−1 (Equ. 18)
The model ensures that the worm numbers are decreasing, if worm,t <0. This process is stopped, when the energy load of settable solids per worm (E settable solids,t ) is sufficient to ensure worm metabolism ( worm,t ≥0). Additionally, the
54 Chapter 2 model assumes the dead worms to be consumed by the living worms. So to tal energy load per day can be described as follows: = + × × × Energyload Total t EsettableSo lids ,t M N ,worm ,t Wworm ,t cDM ,worm E worm (Equ. 19)
Data for c DM, worm (0.14 ± 0.002 g kg 1) and E worm (15.0 ± 0.8 KJ g 1) were de rived from literature data (Bischoff, 2003). Cannibalism is described for this worm species (Hartmann Schröder, 1996) and therefore has to be considered in the model. This option only has to be taken into account, when the natural mortality is lower than the mortality by cannibalism (M C,worm,t ). It appears to be unlikely, that the worms exhibit cannibalism when there are enough dead individuals available. If the
MN,worm,t = 0, it is replaced in Equ.19 by M C,worm,t at the same rate (1.3% d 1).
Module Macroalgae Assimilation TAN (= total ammonia nitrogen) uptake by macroalgae was determined by the daily amount of TAN in the macroalgae tank (B TAN,t ) of the MARE system.
The uptake rates of TAN by macroalgae (r TAN,algae,t ) can be described following Michaelis Menten kinetics: × rmax, TAN ,a lg ae BTAN ,t r = (Equ. 20) TAN ,a lg ae ,t + K s BTAN ,t
where r max,TAN,algae is the maximum uptake rate, B TAN,t the amount of TAN available in the macroalgae tank and K s the half saturation constant. Rela tive uptake of TAN in relation to B TAN,t can be described by the same kinetics: × rmax, TAN %, a lg ae BTAN ,t 100 r = ⋅ (Equ. 21) TAN %, a lg ae ,t + K s BTAN ,t BTAN ,t
55 Chapter 2
BTAN,t can be calculated by using the average TAN concentration in the sys tem water (c TAN,t ) and the flow rate (R) through the macroalgae tank: = × BTAN ,t cTAN ,t R (Equ. 22) The yield (biomass increase in the macroalgae tank) is described by N assimilation of macroalgae determined by the amount of nitrogen in the macroalgae biomass (c N,algae ):
= × 100 × 100 Ya lg ae ,t rt,TAN ,alg ae (Equ. 23) cN,a lg ae cDW ,a lg ae
where c N,algae is the content of N in the algae biomass and C DW,algae the dry weight of the algae. For the modelling the used values were for c Nalgae 6% N
1 DW and for c DW 14% of the fresh weight. The uptake of phosphorus was estimated via the TAN uptake with respect to a Redfield ratio of 16. In consideration of the molar weight of the atoms the ratio is 7.3, so it is r r = t,TAN ,a lg ae (Equ. 24) t,P,a lg ae 7.3
Module Nitrification In contrast to conventional recirculation systems the MARE system was op erated without any additional nitrifying biofilter. However, nitrification was performed because no significant increase in TAN or nitrite was observed. This ammonia converting process plays an important role, because on the one hand, it competes with the ammonia uptake of the macroalgae tank. On the other hand these two processes may be cumulative, leading to an effi cient ammonia removal. The sediment of the detritivorous reactor increased the biologically active surface probably enhancing settlement of nitrifying
56 Chapter 2 bacteria. Thus, the integration of an additional nitrifying biofilter was not necessary. Experimental results of the performance of a fluized bed biofilter from a technical recirculation system were used for modelling the daily nitrifying activity of these additional “nitrifying surfaces” (Fig. 3). Temperature, salinity and turbidity (factors which could influence nitrification) were comparable in both systems. 4,5
4,0 Fig. 3 Performance of a nitrifying fluized bed
] biofilter of a closed recirculation system, stocked -2
m 3,5
-1 with Sea Bass ( Dicentrarchus labrax ), in relation to the inflow ammonia concentration (Wecker and 3,0 Orellana, unpub.). The nitrifying activity was 2,5 determined by the concentration difference of inflow and outflow. It only considers the oxida 2,0
TAN[mg oxidation h tion from ammonia to nitrate and is given by the 1,5 slope (s=2,68; r²=0,86). These data do not include data of zero order ki 1,0 0,4 0,6 0,8 1,0 1,2 1,4 1,6 netic, but ammonia values in the system did not reach more than 1mgL 1 TAN in average in the -3 TAN concentration [mg N dm ] MARE system. For simplifying reasons the model assumed ammonia to be completely con verted to nitrate. This was reasonable, because nitrite concentrations were rather low during the experimental period in the MARE system. Via the slope (s) from Fig. 3 nitrification was calculated as follows: = × rnitrificat ion ,t s cTAN ,t (Equ. 25) where c TAN,t is total ammonia nitrogen (TAN) concentrations of the MARE system. It is were calculated by:
excretion + (c − ×V − r − − r − ) c = TAN ,t TAN ,t 1 nitrificat ion ,t 1 TAN ,a lg ae ,t 1 (Equ. 26) TAN ,t V
57 Chapter 2 where c TAN,t 1 is the total ammonia nitrogen (TAN) concentration in the MARE system from the previous day, V is MARE system volume (4500l) and rnitrification,t 1 is the nitrifying activity of bacteria potentially attached to walls and tubes from the previous day and r TAN,algae,t 1 is the uptake rate of TAN by macroalgae from the previous day. Total surface of MARE was calculated with 60m², (30m² walls and tubes, 30m² sediment). Experimental results of ammonia oxidation rates within the sediment yielded 1,78±0,77g d 1 (Bischoff, unpub.).
Module Denitrification Denitrification is an anaerobic process and plays an important role for the nitrogen removal from the recirculation system. Within the MARE system anaerobic conditions are expected to be prevalent in the lower sediment cen timetres of the detritivorous reactor. Denitrification can be calculated based on of the amount of organic matter
(B OM,t ) left by the worms for denitrifying bacteria. Assuming that no organic matter is accumulating (experimental data: organic matter of sediment dur ing MARE experiment: 2.4 ± 0.5%), B t,OM is calculated on a daily base by: = − × BOM ,t SettableSo lids OM ,t Yworm , cOM ,worm (Equ. 27)
Furthermore, for practical reasons it was assumed that the bacteria are util izing the complete amount of B t,OM for the reduction of nitrate to N 2. Thus, denitrification rates were calculated as follows: = ⋅ rdenitirfic ation ,t z Bt,OM if c nitrate,t ≥ 0 (Equ. 28) where z is the ratio between organic matter needed for the reduction of 1mg nitrate N. OM:N ratio was calculated to be 3.
58 Chapter 2
The nitrate concentration of the recirculation water in the MARE system was calculated by:
c − ⋅V + 98.0 ⋅ r − − r − c = t ,1 nitrate t ,1 nitrificat ion t ,1 denitrific ation (Equ. 29) t,nitrate V
Module Bacterial Assimilation During nitrification and denitrification bacteria are assimilating N and P for biomass production. Heterotrophic denitrifying bacteria can produce more biomass than autotrophic nitrifying bacteria, because energy yield per mol of electron donator (NH 4+, NO 2 ) from nitrification is relatively low (Rheinheimer et al., 1988). Utilization of ammonia nitrogen by Nitrosomonas provides 0,015mg organic N per mg utilized N, while oxidation of nitrite N by Nitrobac- ter gives 0,005mg organic N (Rheinheimer et al. , 1988). During denitrification 0,065mg organic N is formed by the complete reduction of 1mg nitrate N with methanol as hydrogen donator (McCarty et al. , 1969). So N assimila tion by bacteria can be calculated as follows: = × + × rN ,bacteria ,t .0 020 rt,nitrificat ion .0 065 rt,denitrific ation (Equ. 30) For the calculation of P assimilation the stoichiometric ratio between P and
N in bacterial biomass was used. The empirical formula C 250 H611 077 N55 P6S describes the average chemical composition of bacteria, independent from the species (Rheinheimer et al. , 1988). Thus, the stoichiometric N:P ratio is 9.2. Considering the molar mass of the atoms, the weight ratio is given by 4.2. So P Assimilation is calculated by: r r = N ,bacteria ,t (Equ. 31) P,bacteria ,t 4.2
59 Chapter 2
Combination of all modules – simulation of the MARE-system Modelling the nutrient budget in the MARE system was performed with the objective to understand the biological processes and crosslinks in the ex perimental recirculation system. All presented modules individually showed a good correlation between experimental and modelled data. So it can be supposed that all essential processes were adequately modelled. In a simula tion of the complete model a coupling of all stated modules (Fig. 2) was tested. Nitrate and phosphate concentrations are chosen to be the monitored values because of the following reasons: (1) they are simple to analyse, (2) they show very low daily variations, (3) concentration changes happen very slowly and (4) they are influenced by all individual modules.
Nitrate concentrations were calculated including nitrification, denitrification and bacterial assimilation:
c − ⋅V + 98.0 ⋅ r − − r − c = t ,1 nitrate t ,1 nitrificat ion t ,1 denitrific ation (Equ. 32) t,nitrate V Phosphate concentrations can be tolerated by fish at high concentrations, so removal of phosphate plays a minor role in recirculation systems. However, the removal of phosphate must be taken into account, because over a longer period it will accumulate in the system and may result in a water exchange and can be mainly found in unsoluble faeces and excretory products. Con sidering phosphate uptake by algae and bacteria, phosphate concentrations were calculated by:
excretion + faecesUnso lub le + (c − ⋅V − r − − r − ) c = P,t P,t P,t 1 P,bacteria ,t 1 P,a lg ae ,t 1 (Equ. 33) t,P V
60 Chapter 2 2.3 Results 2.3.1 Feasibility of the MARE-system
Water parameters During the 5 months MARE experiment primary water quality parameters were always within safe limits for all cultured organisms: concentrations of toxic substances like nitrite and ammonia were at tolerable concentrations (0.36 ± 0.33 mg L 1 for ammonia and 0.009 ± 0.12 mg L 1 for nitrite during the entire experimental period (11.11.2004 – 1.6.2005). Nitrate was decreas ing from 49.40 mgL 1 to less than 1.45 ±0.03 mg L 1, respectively. pH was stable at 7.97 ± 0.24 but showed a slight decrease at the very end of the ex perimental period. However, pH did never reach critical values below 7. Oxy gen levels averaged 5.73 ± 1.20 mg L 1 during the entire experimental period.
Additionally, MARE fulfilled the demands of a closed recirculation system (less than 10% water exchange per day) (Losordo et al. , 1999) The daily water replacement rate due to evaporation, foam fractionation and losses by main tenance or biomass determination averaged 0.8% of the total system volume (Fig. 4a). No water was exchanged during the entire recirculation system op eration period from April 2004 to December 2005.
200 9,0 a b
150 8,5
100 8,0 pH pH
50 7,5
0 7,0 water consumption [% system volume] system [% consumption water 200 250 300 350 400 200 250 300 350 400 production days production days Fig. 4 (a) Presentation of the cumulative water replacement of the MARE system during the experi mental phase. The daily replacement results from the slope of linear regression, which was 0.8% per day. (r² = 0.996). (b) Development of pH value during the MARE trial.
61 Chapter 2
Fish growth Fig. 5 shows the growth of Sparus aurata in the MARE system over the ex perimental phase of 400 days (data shown from Tank 1). The experimental period started at after Sparus aurata was already cultured for 182 days in the tank showing weights of 71 ± 12g (Tab. 3). Table 3 shows the data from the biomass determinations during the MARE trial. For the numeric model ling experimental data of fish weight (Equ. 1) were taken from fish tank 1. Average feed conversion ratio (FCR) for tank 1 was 1.09 ± 0.32 and ranged from 0.71 to 1.64.
500 0,2 0,4
350
300 400 250 200
0,2 0,4 150
250 300 200 0,2 0,4 150
250 100 0,2 0,4 200 200 50 0,5 1,0 250 150 100 250 200 50 average individual fishweight [g] fishweight individual average 200 150 0 150 100 100 100 50 50 0 0
0
0 100 200 300 400 production days Fig. 5 Growth of Sparus aurata over a cultivation period of 400 days (tank 1). Average individual fishweights are shown in relation to time. Size distributions of Sparus aurata are indicated for the = + ⋅ 5,0 black spots. Growth is described by ln Gt a b t . Regression analysis gave a=0.840; b=0.245; r²=1.00.
62 Chapter 2
Tab. 3 Average fresh weight, total length (TL) and stocking densities of Sparus aurata cultured in the MARE system during the experimental phase per m³ tank volume. A subsample of at least 20 fishes from each tank was taken for measurements. Tank 1 Tank 2 date produc av. av. tion weight av. TL stocking weight TL stocking day [g] [mm] [kg m ³] [g] [mm] [kg m ³]
11.11.04 182 71±12 160±10 3.6 62±10 155±12 3.2
01.12.04 202 73±12 161±8 3.7 78±12 167±9 4.0
22.12.04 223 88±13 174±8 4.5 90±13 179±7 4.6
14.01.04 246 105±15 185±9 5.4 104±15 186±10 5.3
11.02.05 274 133±16 200±8 6.8 132±16 201±10 6.7
14.03.05 305 162±24 213±10 8.3 159±17 211±9 8.1
13.04.05 335 206±27 226±11 10.5 190±15 222±12 9.7
24.05.05 376 268±36 244±11 13.7 242±10 242±11 12.4
16.06.05 399 310±48 259±13 15.8 280±10 251±11 14.3
2.3.2 Modelling the nutrient budget
VALIDATION OF SINGLE MODULES The nutrient budget plays an important role for comprehensive analysis of the system and potential future upscaling. For size optimization of the single modules with respect to the biological and physico chemical interactions within the system a model was developed in order to be able to trace the pathways of the nutrients provided by the feed.
Module Fish a) fish growth and feeding rate Experimental data for feeding rate, absolute and relative growth of fish and FCR are given in Fig. 6a d and show a good correlation to model data (solid line).
63 Chapter 2
The experimental feeding rate was below the recommendation of the feed producer (Fig. 6a, dashed line). However, absolute and relative growth of the fish followed the model predictions (Fig. 6b,c): with increasing fishweight ab solute growth rate is increasing. However, relative growth is decreasing be cause the specific metabolic rate of the fish is also decreasing with increas ing weight (Wehner and Gehring, 1995). Therefore, the proportion of feed used for biomass synthesis also decreases. Consequently, the feed conver sion ratio must be increasing with increasing fish weight (Fig. 6d). Experi mental data were in accordance to model predictions; the model can there fore be considered as suitable for the derivation of these physiological inter actions.
3,0 3,0 a b ] ]
2,5 -1 2,5 -1 fish fish 2,0 -1 2,0 -1
1,5 1,5
1,0 1,0 feeding rate [g d [g rate feeding
0,5 d [g yield absolute 0,5
0,0 0,0 0 50 100 150 200 250 300 0 50 100 150 200 250 300
individual fishweight [g] individual fishweight [g]
0,05 2,0 c d
] 1,8 -1 0,04 1,6 fish -1 0,03 1,4
FCR 1,2 0,02 1,0
0,01 0,8 relative growth [d growth relative 0,6 0,00 0 50 100 150 200 250 300 0 50 100 150 200 250 300 individual fishweight [g] individual fishweight [g] Fig. 6 Validation of the numeric modelling (line) with experimental data from fish biomass determina tions (symbols). The dashed line in (a) indicate the recommended feeding rate from the fish feed pro ducer (feeding rate at 20°C). r² for (a) 0.98; (b) 0.96; (c) 0.95; (d) 0.86.
64 Chapter 2 b) Fractions of solid faeces, excretion and retention The proportion of N and P retained in the fish biomass and the proportion of N and P released to the system environment is of major importance for the evaluation of the system concerning fish yield. Fig. 7 shows the modelled, individual nutrient balance per fish in relation to fish weight. According to physiological principles, the feed proportion used for biomass synthesis de creases with increasing fish weight. This leads to a percentage decrease of feed remaining in the fish body. Consequently, the proportion of excreted substances (e.g. urea, ammonia resulting from protein metabolism) is in creasing. Accordingly, the proportion of excreted solid nutrients is also in creasing. The relative contributions of the different nutrients to the excretory products vary. Nitrogen is mainly released to the water by excretion processes (urea, ammonia) whereas phosphate is a compound of fish faeces (settable solids, Fig. 7a,b). Relative proportions of excreted nutrients in relation to total amount of feed nutrients varied over the experimental period according to the changing utilisation efficiency: for example N excretion rates increased from appr. 50% N d 1 nutrient uptake from feed (at 50g fish weight) to appr. 73% N d 1 nutri ent uptake from feed (at 400g fish weight) (Fig. 7a).
400 75 a b
300 retention fish retention fish faeces undiss. 50 faeces undiss. -1 faeces diss. -1 faeces diss. 200 excretion excretion mg Pd mg mg N d N mg 25 100
0 0 50 100 150 200 250 300 350 400 50 100 150 200 250 300 350 400
fishweight [g] fishweight [g]
Fig. 7 Modelled nutrient balances for fish in the MARE system for Nitrogen (a) and Phosphorus (b). The height of the bars corresponds to the feeding rate of each nutrient. The relative fractions are changing with increasing fish weight because of a different utilisation efficiency.
65 Chapter 2
Experimental data for validation of the model are only available for nitrogen. Fig. 8 shows a good agreement between modelled and experimental data for feeding rate (Fig. 6a), feed retention and excretion (Fig. 7a,b). Experimental data for excretion are derived from the difference of TAN con centrations between inflow and outflow samples of the fishtanks. Raw data of TAN concentrations were expected to be underestimated, because samples were taken before the feeding, but peaks of TAN concentrations occur at least 8 hours after feeding due to diurnal variations of TAN concentrations. 24 hour experiments yielded the factor 2.5 to equalize this underestimation. Therefore, data were corrected with this value. Thus, all model data except the particulate fraction can be validated by experimental values (Fig. 7).
100 200
a ] b -1 ]
-1 80 fish
150 fish -1
-1 60 100 40
50 20 retention [mg N d N [mg retention excretion [mg TAN d [mg excretion
0 0 50 100 150 200 250 300 50 100 150 200 250 300
average fishweight [g] average fishweight [g]
Fig. 8 Validation of the model for nitrogen (line) by available experimental data (symbols): (a) retention (= feeding rate x Y fish,t x fraction in fish, Tab. 1) r²=0,96; (b) excretion (= concentration difference of TAN between inflow and outflow x flow rate x 2,5); r²=0,59. Values for fish weight were summarized in 10g weight classes; TAN concentrations are average values of all experimental data from that time period where each weight class was measured (n>5). It was not possible to validate the model data for phosphorus, because the sensitivity of the orthophosphate analysis for dissolved nutrients was not sensitive enough to measure a difference between inflow and outflow concen trations of the fishtanks. Thus, the applicability of the model for predicting phosphate concentrations can only be tested over the entire experimental MARE phase (see subection 2.2.3.2).
66 Chapter 2
Module foam fractionation Foam fractionation removes all suspended particles from the process water. Data from the rinsing water of the foam fractionators were used to determine the foam fractionation efficiency by regression analysis. Maximum values for r² were calculated by using Microsoft Excel Add In “Solver”: 9,5% of the feed entry (DW) can be found as suspended solids in the MARE system (Fig. 9). As a result the fraction of settable faeces is 15.8% of feed entry (DW), calcu lated by Equ. 13.
0,8 0,6 a b 0,5 0,6 faeces sedimentation
-1 faeces undiss. -1 0,4 foam fractionation suspended solids fish fish -1 0,4 foam fractionation -1 0,3
0,2 g DW d DW g d DW g 0,2 0,1
0,0 0,0 50 100 150 200 100 200 300 400
fishweight [g] production days
Fig. 9 (a) Modelled total amount of faeces (Equ.7, solid line), fractions of undissolved faeces (Equ. 9, dashed line) and suspended solids (Equ. 12, spotted line) in relation to fish weight. Experimental val ues (symbols) from the rinsing water of foam fractionation show a good correlation (r² = 0,71). (b) Calculated daily amounts of faeces transferred to the detritivorous reactor (sedimentation) and the foam fractionators for the entire production period.
Module Detritivorous reactor a) Worm growth and biomass development The daily amount of settable solids (sedimentation from Fig. 9b) is trans ferred to the detritivorous reactor for utilization by Nereis diversicolor. Fig. 10a indicates that the growth of Nereis can be well described by Equ. 13 (p<0,001) (data from Bischoff, 2003). Thus, the energy content of the faeces
(B E) is an adequate variable to describe the growth performance of the worms. Gross energy content of the feed in MARE is given as 20.5MJ kg 1 by the fish feed producer, digestible energy content was 17MJ kg 1 (given by the feed producer), hence the non digestible proportion was 3.5MJ kg 1. Accord
67 Chapter 2 ing to Equ. 9 most of the non digestible proportion was detected as undis solved faeces and the estimated energy content of the faeces was 15KJ g 1 DW. This value could be confirmed by calorific value measurements (14.6 ± 0.5 KJ g DW, data not shown) of the faeces. The individual energy load is influenced by the worm number living in the sediment (Equ. 14). The bioreactor was stocked with appr. 2000 individuals of Nereis at day 36 with a total biomass of 1.8 ± 0.5kg. The average individ ual worm weight was 890 ± 260mg (n= 270). Biomass determination at the start of the MARE experimental phase (day 182) resulted in worm numbers approximating 50.000 with a total biomass of 4.4 ±3.7kg (Fig. 10b). Individ ual worm weight averaged 87 ± 73mg. This clearly indicates a reproduction event between day 36 and 182 (data not shown). After spawning Nereis di- versicolor dies (monotelic reproduction). The high standard deviation values in the individual weight (± 84%) at the beginning of the MARE experiment indicates that there were still larger Nereis individuals which had not yet spawned. Biomass and total number of worms in the bioreactor showed a continuous decrease during the experimental period (Fig. 10b, d). The aver age worm weight was nearly constant until day 340 and increased at the end of the experimental phase (Fig. 10c).
68 Chapter 2
3 12x10 3
a 3 b 2 10x10
] 3 -1 8x10 1 6x10 3 0 4x10 3
-1 [g] biomass 2x10 3 growth rate [% d [% rate growth
-2 0
-3 -2x10 3 0,0 0,5 1,0 1,5 2,0 2,5 150 200 250 300 350 400
energy content of solids [KJ d -1 worm -1 ] production days
1,0 60x10 3 c d 50x10 3 0,8 40x10 3 0,6 30x10 3 0,4 20x10 3 average weight [g] weight average individual numbers individual 0,2 10x10 3
0,0 0 150 200 250 300 350 400 150 200 250 300 350 400
production days production days
Fig. 10 (a) Growth performance of Nereis diversicolor in relation to energy content of the solids (Bisch off, 2003). Regression analysis gave: max = 2.78; min = 2.72; K s = 0.22. (b d) Modelled and experimen tal data of (a) total biomass, (b) average worm weight and (c) individual numbers of Nereis diversicolor in the detritivorous reactor in the MARE system. Dashed lines are calculations without cannibalism, solid line includes a relative mortality of 1.3% d 1. Validation with experimental data gave for (a) r² = 0.98 and (b) r² = 0.93 and (c) r² = 0.99 and (d) r² = 0.97. Considering elevated energy input by dead worms (Equ. 19) and potential cannibalism under certain conditions, model data show a good correlation to the experimental data. By using Equ. 14 relative and absolute worm growth rates can be calculated for the experimental period. The results indicate a feed undersupply of the worms because of the elevated numbers of individu als due to the reproduction event before the start of the experimental period. The energy load provided by the feed (fish faeces) turned out to be not suffi cient to satisfy the basic metabolic needs (Fig. 11a). This may explain the negative absolute and relative growth rates until experimental day 340 (Fig.
69 Chapter 2
11b,c), and the increase afterwards, resulting in the increased average worm weight at the end (Fig. 10c).
1,6 3 a b ]
-1 2
1,2 ] -1
worm 1 -1
0,8 0
-1
0,4 d [% rate growth -2 energy load [KJ d [KJ load energy
0,0 -3 200 250 300 350 400 200 250 300 350 400 production days production days
40 c 20
0 ]
-1 -20 -40 -60 yield [g d [g yield -80 -100
-120 200 250 300 350 400 production days Fig. 11 Modelling of the energy load (a), relative growth rate (b) and absolute growth rate (c) in the detritivorous reactor during the complete MARE experiment. The model was calculated without mor tality by cannibalism (dashed line) and regarding a relative mortality of 1.3% d 1 (solid line).
Module Macroalgae Filter Ammonia uptake by macroalgae is described by a Michaelis Menten kinetic (Equ. 20). Experimental values show a better correlation to Equ. 21 (Fig. 12b, dashed line, r² = 0.75), caused by a better adaptation to low nutrient concentrations in contrast to the absolute uptake rates according to Equ. 20 (solid line). Absolute uptake rates were used for model calculations, because an underestimation of maximum uptake rate in Equ. 21 was expected (r max = 0,06g TAN h 1 m 2, Fig. 9a).
70 Chapter 2
Fig. 12c compares modelled and experimental data of ammonia uptake dur ing the MARE trial, which did not show a good correlation due to daily varia tions of the TAN concentrations and increasing modelled TAN concentra tions.
0,30 a b ]
-2 0,25 100 m -1 0,20 80
0,15 60
0,10 40 TAN uptakee [% TAN load] [% TAN uptakee TAN
ammonium uptake [g TAN h [g TAN uptake ammonium 0,05 20
0,00 0 0,0 0,5 1,0 1,5 2,0 0,0 0,5 1,0 1,5 2,0
TAN load [g TAN h -1 m -2 ] TAN load [g TAN h -1 m -2 ]
0,30 500 c d ]
-2 0,25 400 m -1
0,20 ] -1 300
0,15
200
0,10 yield [g FW d
100
ammonium uptake [g TAN h [g TAN uptake ammonium 0,05
0,00 0 200 250 300 350 400 150 200 250 300
production days production days
Fig. 12 Macroalgae filter: (a) absolute TAN uptake in relation to TAN load. Numeric analysis gave for Equ. 26: r max = 0.11; K s = 0.25; r² = 0.28; (b) relative TAN uptake in relation of TAN load. Solid line represents zero order kinetic from Fig. 9a (r² = 0,64), dashed line from Equ. 25. Regression analysis gave: r max = 0.06; K s = 0.05; r² = 0,75. c) modelled and experimental data of ammonia uptake during the MARE trial. Experimental data are representing the concentration difference between inflow and outflow of the macroalgae tank. d) Modelled and experimental growth yield in the macroalgae tank during the experimental time.
In Table 4 the growth yield of Solieria chordalis in the MARE system is shown. At experimental day 125 (15.09.2004) Solieria was stocked for the first time. Recommended stocking density was 8kg FW m 2 surface area (Sylter Algenfarm, pers. comm.). Therefore, for the MARE tank with a surface area of 2.3m² a total biomass of 18kg FW was needed. Technical problems
71 Chapter 2
(delivery bottlenecks of the algae farm) resulted in a total stock of only 10kg FW and 11.2kg FW, respectively. Tab. 4 Start and end stockings of Solieria chordalis in the MARE system during the ex perimental trial. Numbers present fresh weight. start stock- end stock- specific experimental pe- growth rate days ing density ing density growth rate riod FW d -1 m-2 [g] [g] [100 d -1]
Solieria chordalis I first stocking
15.09.04 – 29.09.04 13 10000 11806 60 1.28
29.09.04 – 06.10.04 6 11806 13653 134 2.42
06.10.04 – 13.10.04 6 13653 14561 66 1.07
13.10.04 – 21.10.04 7 14561 15454 56 0.85
21.10.04 – 27.10.04 5 15454 16039 51 0.74
27.10.04 – 10.11.04 13 16039 16800 26 0.36
10.11.04 – 17.11.04 6 16800 17000 15 0.20
17.11.04 – 24.11.04 6 15600 15731 10 0.14
24.11.04 – 10.01.05 46 14331 15000 6 0.10
Solieria chordalis II second stocking
10.01.05 – 21.01.05 10 11195 11836 28 0.56
21.01.05 – 02.02.05 11 11836 15649 151 2.54
02.02.05 – 09.02.05 6 15649 17201 113 1.58
09.02.05 – 18.02.05 8 17201 20069 156 1.93
18.02.05 – 25.02.05 6 18000 18453 33 0.41
25.02.05 – 03.03.05 5 18000 18785 68 0.85
03.03.05 – 11.03.05 7 18000 20758 171 2.04
11.03.05 – 17.03.05 5 18000 18682 59 0.74
17.03.05 – 30.03.05 12 18000 21054 111 1.31 During the first experimental phase ( Solieria chordalis I) recommended stocking densities were not reached due to decreasing growth rates. There fore, a second stocking ( Solieria chordalis II ) was necessary. Additionally, a heavy fouling of epiphytes could be observed and the Solieria thalli were in 72 Chapter 2 terspersed with the green algae Enteromorpha sp.. Because of delivery bot tlenecks the second stocking could not be performed before 10 th January 2005. During the second experimental phase growth rates increased, but strong fouling of epiphytes and Enteromorpha growth occurred again after a few weeks.
COMBINATION OF ALL MODULES – SIMULATION OF THE MARE SYSTEM The prediction of nitrate and phosphate concentrations in the research recir culation system needs a comprehensive understanding of all biological proc esses in the MARE system. The presented modules were combined according to Equ. 29 und 30. Modelled and experimental nitrate and phosphate con centrations are shown in Fig. 13a,b.
100 40 a b
80 30 ] -3 ] -3 60 20
40
Nitrate [mgNitrate N dm 10 Phosphate [mgPhosphate dm P 20
0 0
200 250 300 350 400 200 250 300 350 400 production days production days
50 ] 30 -3 ] -3 c d 25 40
20 30
15
20 10
10 5 Modelled nitrate concentrationnitrate [mg dm P Modelled
0 P [mg dm concentration phosphate modelled 0 0 10 20 30 40 50 0 5 10 15 20 25 30
experimental nitrate concentration [mg N dm -3 ] experimental phosphat concentration [mg P dm -3 ] Fig. 13 Experimental (symbols) and modelled data (solid line) of (a) nitrate concentrations (b) phos phate concentrations in the process water of MARE over the experimental period. (c d) linear regres sion analysis: Modelled vs. experimental nutrient concentrations (c) y = 1.05x + 1.38; r² = 0.84 (d) y = 0.92x + 2.04; r² = 0.87
73 Chapter 2
Linear regression analysis (Fig. 14c,d) underlines the suitable predictions of the model over the entire experimental period. Standard error of regression was 5.7mg N dm ³ for nitrate and 1.8mg P dm 3 for phosphate.
The modelled N:P ratio was compared to the experimental data (Fig. 14) and also showed a very good correlation. Both results underline that the MARE model describes the process in the experimental system adequately.
20 14 a b 12
15 10
8 10
N:P ratio N:P 6
5 ratio N:P Modelled 4
2 0
0 200 250 300 350 400 0 2 4 6 8 10 12 14 production days experimental N:P ratio Fig. 14 (a) Modelled (solid line) and experimental data (symbols) of the stoichiometric N:P ratio in the process water over the entire experimental period (b) linear regression analysis: Modelled vs. Experi mental stoichiometric N:P ratio: y= 1.04x + 0.30, r²=0.87.
2.4 Discussion The major aims of the present study were: a) the operation of the integrated system within safe limits for all cultured organisms (e.g. keeping ammonia and nitrite concentrations low, maintain pH values and oxygen saturation at constant levels, stabilizing nitrate and phosphate concentrations or even re duce them through culture of suitable secondary organisms). b) to test the feasibility of recycling nutrients within the different components of the sys tem. c) achieve a long experimental period to obtain sufficient data for mod elling. The different aims are discussed in detail in the following chapters.
74 Chapter 2
2.4.1 Feasibility of the MARE system The primary aim of this study was to ensure a safe operation of the system. Compared to conventional recirculating aquaculture systems the stock den sity of fish was low: at the end of the trial it was 7kg m 3 based on the sys tem volume, 15kg m 3 based on the fish tank volume. Although the concen trations of inorganic dissolved nutrients (ammonia and nitrite) as well as oxygen levels would have allowed a biomass increase, the volume of the fish tanks limited the numbers of fish regarding the welfare of the animals. Lar ger tanks could not be used due to the localisation and technical features of the laboratory. Fish grew well in the system, they reached commercial size after 14 15 months. The reduced growth compared to other systems (12 13 months) (Lupatsch and Kissil, 1998; Lupatsch et al., 2003; Colloca and Cerasi, 2005; Fig. 16) was supposed to be caused by lower temperatures, a reduced feed ing rate, irregular feeding before and on biomass determination days, larger maintenance/construction work or foam fractionator failures, which oc curred especially at the beginning of the experiment. However, in compari son to cultivation in net cages observed growth was faster. Fish culture in net cages resulted in rearing periods of 15 16 months due to lower water temperatures surrounding the cages (Porter et al. , 1986; Colloca and Cerasi, 2005). Noticeable, food conversion ratio was very good (1.0 – 1.3) during the com plete experimental period and in the range described for commercial farms (Colloca and Cerasi, 2005). The measured value of 1.1 is below the FCR (1.5) described by Lupatsch and Kissil (1998, Fig 16b) and Porter et al. (1986) FCR = 2.7. This might be due to enhanced efforts of fish feed producers in the last few years to adapt feed compositions better to the requirements of the cultured species. To ensure safe system operation the control of the faeces played a key role, due to the potential of influencing water quality severely (Cripps and Berg heim, 2000). MARE included a two step solid separation (sedimentation and
75 Chapter 2 foam fractionation) to ensure best water quality (Losordo, 1999; Summerfelt, 2002). In contrast to conventional systems the particulate waste (fish faeces and remaining food pellets) was not eliminated from the system, conse quently leaching occurred within the first 6 hours (Lupatsch and Kissil, 1998) and could not be influenced. However, concentrations of toxic nutri ents and oxygen saturation did not reach critical values in the MARE system at any time due to efficient macroalgae filter. The decrease of pH at the end of the experimental period can be explained due to the lower growth performance of the macroalgae (further discussed in subsection 2.4.2). Because of the reduced growth rates of macroalgae the uptake of essential CO 2 for photoautotrophic production resulting from the fish metabolism was not efficient enough. Consequently, carbon dioxide
(CO 2) can react in the water to form bicarbonate ions (HCO 3 ) as well as car bonate ions (CO 32 ), depending on the pH value of the water. 2.4.2 Nutrient recycling by integration of secondary organisms (Solieria, Nereis ) Second aim of the study was to investigate the feasibility of an enhanced nu trient recycling by integration of biological secondary steps ( Solieria chordata, Nereis diversicolor ). Fig. 17 shows the modelled daily retention of N and P for all cultured organisms during the entire experimental period in the MARE system.
60 80 a Sparus aurata b Sparus aurata 50 Solieria chordalis 70 Solieria chordalis Nereis diversicolor 60 Nereis diversicolor ] 40 bacteria ] bacteria -1 -1 50 30 40 20 30 10 20 10 P retention [% d [% Pretention N retention [% d [% retention N 0 0 -10 -10 -20 -20 200 250 300 350 400 200 250 300 350 400
production days production days
Fig. 15 Daily retention of (a) nitrogen and (b) phosphorus of the nutrient intake by feed for all culti vated organisms in the MARE system during the entire experimental period.
76 Chapter 2 a) Fish unit As expected, nutrient retention in fish decreased over the cultivation period due to decreased feed utilization with increasing fish size (Lupatsch and Kis sil, 1998, Fig. 15a, b). In this experiment the determined values for N reten tion in fish (31±3%) are within values described in the literature (11 36%, Hargreaves 1998; Paspatis et al. , 2000; Van Ham et al. , 2003). For P reten tion in fish (45±5%) values obtained from experiments are higher than de scribed in literature (21 29%; Krom et al. , 1985; Porter et al. , 1987b; Krom and Neori, 1989; Krom et al. 1995). This can be caused by the selected kind of feed, as relative P rentention depends on type, concentration, availability and digestibility of this nutrient, which can vary due to different sources (fishmeal, fishoil, crops etc.) (Hua and Bureau, 2005). b) Macroalgae unit Around 17% of N and 13% of P were additionally retained in macroalgae biomass (Tab. 6). The macroalgae filter was sufficient to ensure low concen trations of nitrite and ammonia in the recirculation water of the MARE system. However, growth performance of the macroalgae was inadequate and can be explained by the following reasons: concentrations of ammonia were very often only sufficient to cover the minimum physiological requirements of the algae but not enough for significant growth. Additionally, artificial illu mination results in lower nutrient uptake rates compared to natural light. Solieria chordalis did not switch to nitrate uptake because of the preference for ammonia, which was permanently available at low levels in the tank. It takes at least a 12 hour absence of this nutrient to induce nitrate uptake in Soleria chordalis (Lüning, pers. comm.), but such a situation is unlikely to happen in recirculation systems due to a constant availability of ammonia caused by fish metabolism. Although concurrent uptake of ammonia and nitrate is known for macroalgae (Ahn et al., 1998), the measurements clearly showed that Solieria chordalis did not take up nitrate. Biofouling of epiphytes occurred as a consequence of the poor growth and insufficient stock density of macroalgae at the beginning of the experiment (recommended stocking density approx. 18 kg m 2). In conclusion, a smaller macroalgae cultivation
77 Chapter 2 unit might have yielded better results with respect to these conditions. An other explanation for the insufficient performance of the macroalgae unit could be a limitation by micronutrients, although it is assumed that suffi cient amounts of micronutrients are provided to the system by fish feed (Metaxa et al. , 2006). Nevertheless, it might be possible that a negative effect on some essential ions occurred (e.g. oxidizing Fe 3+ to Fe 2+ ) as a result of the utilization of ozone. Thus, some essential ions may not further be available for the macroalgae. Addition of micronutrients was performed for a limited time period, but results showed no better growth performance of the macro algae and addition was therefore terminated. c) Worm unit Caused by the spontaneous reproduction of Nereis diversicolor a biomass loss due to the monotelic behaviour (death of mature worms after the spawn ing incident) appeared. This resulted in negative growth rates of Nereis diver- sicolor since mature individuals possess a higher bodyweight than juveniles. This further leads to a negative nutrient retention. Simultaneously the re production incident increased stock density within the detritivorous tank up to about 50 000 individuals/m². The available energy load per worm de creased far below minimum requirements of an individual (Bischoff, 2003). MARE has to be considered as a simplified ecosystem. A destabilisation of this “artificial ecosystem” due to insufficient energy uptake by the worms and high individual worm numbers did not happen because these processes were regulated by decreased survival of juvenile worms. The number of worms decreased until sufficient energy was provided for each worm, and subsequently growth rates and average weight increased to positive values at the end of the experiment. In later experiments positive growth of worms up to spawning size could be observed (see Chapter 5).
In summary, the nutrient budget gives an overview of the retention of N and P in each module of the system for the entire experimental period (produc tion day 182 – 399) in relation to the feed uptake (Tab. 6).
78 Chapter 2
Tab.6 Nutrient budget for nitrogen (N) and phosphorus (P) according to the modelled data. They are presented as sum of absolute and relative amount from the MARE experiment. Nitrogen Nitrogen Phosphorus Phosphors [g] [%] [g] [%]
Sparus aurata 1194 30 302 43
Nereis diversicolor 50 1 7 1
Solieria chordalis 666 17 92 13
Foam fractionation 145 4 62 9
Bacteria 134 3 32 5
Denitrification 1417 35
Process water 259 6 118 17
Sum 3765 94 599 86
Feed intake 4003 100 701 100
Difference 238 6 102 14 d) Nitrogen recycling It seems obvious that one of the major processes within the system was de nitrification, possibly mainly located in the sediment of the detritivorous re actor: bacterial degradation of the organic matter could create anaerobic conditions in the sediment. Combined with observed nitrate concentrations and organic content in the sediment favourable conditions for denitrification can occur. Nitrate will be reduced to elementary nitrogen during denitrifica tion; ascending gas bubbles could be observed during sediment sampling processes. Another considerable conversion process of nitrate is the bacterial assimila tory nitrate reduction. This process may therefore also be responsible for a certain part of the nitrate removal, but a considerable part was accumulat ing in the system water and in the rinsing water of the foam fractionators. Applied to the model, 94% of the nutrient balance of the system could be ex plained for nitrogen as well as 86% for phosphorus, so the model can be
79 Chapter 2 considered as suitable for the simulation of a marine artificial recirculation system. e) General considerations Generally, it seems that the spatial dimensions of the different biological secondary units did not exactly match in the experiment. Therefore, system optimisation can include stocking density, harvesting processes or water flow rates. In order to calculate the right spatial dimensions for such a sys tem it is necessary to define the criteria for optimisation: e.g. the financial benefit of the secondary steps or the nutrient retention within the different biological secondary elements. It is not possible to optimise all criteria simul taneously. Maximum biomass yield in all components at the same time is also impossible due to the different physiological requirements of the cul tured organisms. Therefore, the model can be a powerful tool for optimising integrated recirculating systems. f) Commercial applications Although nutrient retention of the biological secondary units accounted for only one fifth of the total nutrient budget, the financial benefit of the addi tional biomass production can significantly improve the economical out come. For example, the harvested worms may be directly sold as bait for fishers or as high quality aquaculture food (high proportions of polyunsatu rated fatty acids). Preliminary calculations based on the MARE data and cur rent market prices of €30/kg indicate a cost coverage of one year´s fish feed by selling the additional crop of worms. The direct sale of macroalgae appears to be more difficult, but macroalgae can be also used in aquaculture as feed for more profitable organisms, like abalone ( Haliothis spec.) or as dietary supplement for aquaculture fish spe cies e.g. Sparus aurata . However, results from feeding experiments (this study, data not shown) did not show a significant difference in fish growth with a feed replacement of 5 %. Further options for the commercial usage of
80 Chapter 2 cultured macroalgae are products from fermentation processes of algae (e.g. for cosmetics and pharmacy). 2.4.3 Modelling In aquaculture a broad spectrum of models were developed for different spe cies and various aquaculture productions forms (Roland and Brown, 1990; Piedrahita, 1990; Watten, 1992; Kochba et al. , 1994; Ellner et al. 1996, Cho and Bureau, 1998; Pagand et al. , 2000a; Lefebvre et al. 2001b; Gasca Leyva et al. , 2002). It is obvious, that in models natural processes can only be de scribed in a simplified way. During the development of this model 15 mod ules were identified and characterised within the particular system compo nents (fish tank, macroalgae tank, detritivorous reactor, foam fractionation), although some were overlapping, e.g. nitrification. For practical reasons different assumptions concerning the modelling proc ess were made. The following paragraph lists and explains these assump tions: It was assumed that the excretion of urea by fish is of minor importance. This was done despite the results presented by Dosdat and co authors (1996) which showed that the excretion of urea represents about 15 % of to tal nitrogen excretion. Measurements of urea production during the experi mental period were not performed and therefore modelling of urea produc tion and remineralisation could not be achieved. It was further assumed that all settable solids are transferred to the detritivorous reactor. Another assumption was that Nereis diversicolor could not build up any lipid reserves. Lipids are well recognised as an energy source for marine invertebrates (Luis and Passos, 1995; Graeve and Wehrtmann, 2003; Sewell, 2005). Beyond that, lipids provide structural components for membranes in the form of phospholipids (Sargent and Whittle, 1981). This lack of reserves and essential components led to death of worms during the course of the experiment. Another pre assumption was that no organic matter accumulates in the sediment of the detritivorous reactor. This could be proved by experimental results of sediment samples during the experiment and therefore this assumption was included into the 81 model. Additionally, it Chapter 2 sumption was included into the model. Additionally, it was assumed that all the organic matter not being used by the worms (B t,OM ) will be used for deni trification processes. Conditions favouring denitrification or ANAMMOX (an aerobic ammonia oxidation) (suboxic areas, sufficient contents of organic matter and elevated nitrate concentrations) occurred during the experiment.
The ANAMMOX process leads to the formation of N 2 assuming suboxic/anoxic conditions. Suboxic conditions are considered to develop at organic particle surfaces and in deeper sediment regions. However, elevated nitrite concentrations could not be detected being a prerequisite for the on set of the ANAMMOX process. Therefore, this process may be neglectable. Additionally, nitrate concentrations rose before integration of the macroalgae unit, indicating considerably higher nitrification rates compared to denitrifi cation rates, although no quantifications were made. Although all these processes have to be taken into account, for practical reasons it was as sumed that all available ammonia is converted to nitrate. In contrast to many other static models, this dynamic model refers to the variable “production day” with the changing variable “fish growth” of Sparus aurata. Based on these data the influence of all presented modules was in vestigated. Within a numeric model iterative algorithms were developed for nutrient flows of each production day. It is based on functional correlations resulting from theoretical considerations, literature data and regression analyses of empirical and literature data. The informational value of the model was evaluated by the modelled and ex perimental concentrations of nitrate and phosphate in the MARE system and showed good correlations (Fig. 14). However the model was not yet validated by an independent data set. The empirical data for fish growth were used for development and linear regression analyses for numeric modelling. So, transferability of the results to other recirculation system could not be sta tistically affirmed. Nevertheless, it can be assumed that the model represents a valuable tool to calculate the proper dimensions of future integrated recir culating system and to forecast nutrient flows within the different compo
82 Chapter 2 nents of such a system. It can be used as a reasonable basis for further in vestigations into this area. Despite all achievements of the model systems like the presented MARE sys tem still struggle with problems like the accumulation of dissolved inorganic nutrients. The excess of certain critical values can cause limitations in growth performance or even death of the cultured organisms. So far no rec ommendations concerning harvesting of the cultured organisms were made. In order to avoid production failures (due to e.g. reproduction events) worms for example should be harvested before they start reproduction.
2.5 Conclusions MARE is an innovative concept for new marine aquaculture cultivation sys tems which allows not only water reprocessing but also nutrient recycling with the aim of enhancing economical profitability. The presented numeric modelling identified all essential biological processes and linked them to a complete tool, which is able to predict the growth of the cultured organisms as wells as concentrations of accumulating dissolved inorganic nutrients (ni trate and phosphate). The observed results enable the use of the model for future scientific questions, e.g. to integrate further biological steps for im proving the nutrient recycling to replace technical solutions. The replace ment of foam fractionators by filtering organisms (e.g. bivalves) and the de velopment of a suitable microalgae photobioreactor system (Kube et al. , in prep., see Chapter 3 & 4) may be steps towards this direction. Although the first attempt of this new type of recirculation system was beyond optimal configuration, the experimental data and the developed model can help to calculate the nutrient fluxes and module sizes in virtual simulations to im prove the biological and economical efficiency of future commercial recircula tion systems (von Harlem, 2006).
83 Chapter 2 2.6 Acknowledgements This work was founded by Deutsche Bundesstiftung Umwelt (DBU) and the EU (Interreg IIIA). We thank Thomas Hansen (IFM GEOMAR) for measuring POC and PON. 2.7 References Ahn O., Petrell R. J., and Harrison P. J. (1998). Ammonium and nitrate up take by Laminaria saccharina and Nereocystis luetkeana originating from a salmon sea cage farm. Journal of Applied Phycology 10: 333 – 340. Asgard T., Austreng E., Holmefjord I., and Hillestad M. (1999). Resource effi ciency in the production of various species. In : Svennevig N., Reinertsen N., New H., and New M. (eds). Sustainable aquaculture: food for the future ? A.A. Balkema, Rotterdam, Holland.: 171 183. Barak Y. and van Rijn J. (2000). Biological phosphate removal in a prototype recirculation aquaculture treatment system. Aquacultural Engineering 22: 121 136. Bischoff A.A. (2003). Growth and mortality of the polychaete Nereis diversi- color under experimental rearing conditions. M.Sc.thesis, Institute of Marine Research & Department of Animal Sciences, Chairgroup of Fish Culture and Fisheries, Christian Albrechts University Kiel, Germany/Wageningen Univer sity, The Netherlands; 103pp. Buschmann A.H., Troell M., Kautsky N., and Kautsky L. (1996). Integrated tank cultivation of salmonids and Gracilaria chilensis (Rhodophyta). Hydro biologica 326/327: 75 82. Chamberlain G. and Rosenthal H. (1995). Aquaculture in the next century: Opportunities for growth challenges of sustainability. World Aquaculture 26: 21 25. Cho C. Y. and Bureau D. P. (1998). Development of bioenergetic models and the fish PrFEQ software to estimate production, feeding ration and waste output in aquaculture. Aquat. Living Resour. 11 (4): 199 – 210. Chopin T., Buschman A.H., Halling C., Troell M., Kautsky N., Neori A., Kraemer G.P., Zertuche Gonzales J.A., Yarish C., and Neefus C. (2001). Inte grating seaweeds into marine aquaculture systems: A key toward sustain ability. Journal of Phycology 37: 975 986. Chopin T., Sharp G., Belyea E., Semple R., and Jones D. (1999a). Open wa ter aquaculture of the red algae Chondrus crispus in Prince Edward Island, Canada. Hydrobiologica 398/399: 417 425.
84 Chapter 2
Chopin T., Yarish C., Wilkes R., Belyea E., Lu S., and Mathieson A. (1999b). Developing Porphyra /salmon integrated aquaculture for bioremediation and diversification of the aquaculture industry. Journal Applied Phycology 11: 463 472. Colloca, F. and Cerasi, S. (2005). Cultured Aquatic Species Information Pro gramme Sparus aurata . FAO Inland Water Resources and Aquaculture Ser vice (FIRI). FAO FIGIS. Cripps S.J. and Bergheim A. (2000). Solids management and removal for in tensive land based aquaculture production systems. Aquacultural Engineer ing 22: 33 56. Dosdat A., Servias F., Metailler R., Huelvan C., and Desbruyères E. (1996). Comparison of nitrogenous losses in five teleost fish species. Aquaculture 141: 107 – 127. Ebenhöh (2004). www.icbm.de/~mathmod/pages/lectures/mm/Kap00_Ziele_Bedeutung.pdf. Ellner S., Neori A., Krom M. D., Tsai K., and Easterlling M. R. (1996). Simu lation model of recirculating mariculture with seaweed biofilter: Development and experimental tests of the model. Aquaculture 143 (2): 167 – 184. Gasca Leyva E., Leon C. J., Hernandez J. M., and Vergara J. M. (2002). Bio economic analysis of production location of sea bream ( Sparus aurata ) culti vation. Aquaculture 213: 219 – 232. Graeve M. and Wehrtmann I. S. (2003). Lipid and fatty acid composition of Antarctic shrimp eggs (Decapoda: Caridea). Polar Biology 26: 55 61. Hall P.O.J., Kollberg S., and Samuelsson M.O. (1992). Chemical fluxes and mass balance in a marine fish cage farm. IV.Nitrogen. Marine Ecology Pro gress Series 89: 81 91. Hargreaves J.A. (1998). Nitrogen biogeochemistry of aquaculture ponds. Aquaculture 166: 181 212. Hua K., Liu L., and Bureau D. P. (2005). Determination of phosphorous frac tions in animal protein ingredients. Journal of Agricultural and Food Chem istry 53: 1571 – 1574. Kochba M., Diab S., and Avnimelech Y. (1994). Modelling of nitrogen trans formation in intensively aerated fish ponds. Aquaculture 120 (1 – 2): 95 – 104. Krom M. D., Ellner S., Rijn J. van, and Neori A. (1995). Nitrogen and phos phorous cycling and transformations in a prototype ‘non polluting’ inte grated mariculture system, Eilat, Israel. Marine ecology progress series, 118 (1 – 3): 25 – 36.
85 Chapter 2
Krom M.D. and Neori A. (1989). A total nutrient budget for an experimental intensive fishpond with circularly moving seawater. Aquaculture 83: 345 358. Krom M.D., Porter C., and Gordin H. (1985). Nutrient budget of a marine fish pond in Eilat, Israel. Aquaculture 51: 65 80. Lefebvre S., Bacher C., Meuret, A., and Hussenot J. (2001). Modelling Nitro gen Cycling in a Mariculture Ecosystem as a tool to evaluate its outflow. Es tuarine, Coastal and Shelf Science 52 (3): 305 325. Losordo T.M., Masser M.P., and Rakocy J.E. (1999). Recirculating aquacul ture tank production systems: A review of component options. SRAC Publi cation 453. Luis O. J. and Passos A. M. (1995). Seasonal changes in lipid content and composition of the polychaete Nereis ( Hediste ) diversicolor . Comparative Bio chemistry and Physiology. Part B: Biochemistry & Molecular Biology [COMP. BIOCHEM. PHYSIOL., B]. VOL. IIIB, no. 4: 579 – 586. Lupatsch I. and Kissil G.W. (1998). Predicting aquaculture waste from Gilt head Seabream ( Sparus aurata ) culture using a nutritional approach. Aquat. Living Ressour. 11: 265 268. Lupatsch I., Kissil G. W., and Sklan D. (2003). Comparison of energy protein efficiency among three fish species gilthead sea bream ( Sparus aurata ), European sea bass ( Dicentrarchus labrax ) and white grouper ( Epinephelus aeneus ): energy expenditure for protein and lipid deposition. Aquaculture 225 (1 – 4): 175 – 189. McCarty P. L., Beck L., and Amant P. S. (1969). Biological denitrification of wastewater by addition of organic materials. P 1271. In : Proceedings of the 24 th Indiana Waste Conference, no. 135. Purdue University, Lafayette, Ind. Metaxa E., Deviller G., Pagand P., Alliaume C., Casellas C., and Blancheton J. P. (2006). High rate algal pond treatment for water reuse in a marine fish recirculation system: Water purification and fish health. Aquaculture 252 (1): 92 – 101. Neori A., Cohen I., and Gordin H. (1991). Ulva lactuca biofilters for marine fish pond effluents. II. Growth rate, yield and C:N ratio. Botanica Marina 34: 483 489. Neori A., Shpigel M., and Ben Ezra, D. (2000). A sustainable integrated sys tem for culture of fish, seaweed and abalone. Aquaculture 186: 279 291. Newkirk G. (1996). Sustainable coastal production systems: a model for in tegrating aquaculture and fisheries under community management. Ocean Coast. Manag. 32: 69 83.
86 Chapter 2
Pagand P., Blancheton J. P., and Casellas C. (2000). A model for predicting the quantities of dissolved inorganic nitrogen released in effluents from a sea bass ( Dicentrarchus labrax ) recirculating water system. Aquacultural Engi neering 22 (1 – 2): 137 – 153. Paspatis M., Boujard T., Maragoudaki D., and Kentouri M. (2000). European sea bass growth and N and P loss under different feeding practices. Aquacul ture 184 (1 – 2): 77 – 88. Petrell R.J. and Alie S.Y. (1996). Integrated aquaculture of salmonids and seaweeds in open systems. Hydrobiologica 326/327: 67 73. Pfeiffer T.J. and Rusch K.A. (2000). An integrated system for microalgal and nursery seed clam culture. Aquacultural Engineering 24: 15 31. Porter C. B., Krom M. D., and Gordin H. (1986). The effect of water quality on the growth of Sparus aurata in marine fish ponds. Aquaculture 59: 299 – 315. Porter C. B., Krom M. D., Robbins M. G., Brickell L., and Davidson A. (1987). Ammonia excretion and total N budget for gilthead sea bream ( Sparus au- rata ) and its effect on water quality conditions. Aquaculture 66: 287 – 297. Rheinheimer G., Hegemann W., Raff J. and Sekoulov I. (1988). Stickstoff kreislauf im Wasser: Stickstoffumsetzungen in natürlichen Gewässern, in der Abwasserreinigung und Wasserversorgung. R. Oldenburg Verlag GmbH, München. Roland W. G. and Brown J. R. (1990). Production model for suspended cul ture of the Pacific oyster, Crassostrea gigas . Aquaculture 87: 35 52. Sargent J. R. and Whittle K. J. (1981). Lipids and hydrocarnbons in the ma rine food web. In: Longhurst A. (ed) Analysis of marine ecosystems. Academic Press, London, pp. 491 – 533. Schneider O., Sereti V., Eding E.H., and Verreth J.A.J. (2005). Analysis of nutrient flows in integrated intensive aquaculture systems. Aquacultural Engineering 32: 379 401. Sewell M. S. (2005). Utilization of lipids during early development of the sea urchin Evechinus chloroticus . Marine Ecology Progress Series 304: 133 – 142. Summerfelt S.T. (2002). An integrated approach to aquaculture waste man agement in flowing water systems. Proceedings of the 2nd International Con ference on Recirculating Aquaculture: 87 97. Troell M., Halling C., Neori A., Chopin T., Buschman A.H., Kautsky N., and Yarish C. (2003). Integrated mariculture: asking the right questions. Aqua culture 226: 69 90.
87 Chapter 2
Van Ham E. H., Berntssen M. H., Imsland A. K., Parpoura A. C., Wendelaar Bonga S. E., and Stefansson S. O. (2003). The influence of temperature and ration on growth, feed conversion, body composition and nutrient retention of juvenile turbot ( Scophthalmus maximus ). Aquaculture 217: 547 – 558. Vandermeulen H. and Gordin H. (1990). Ammonium uptake using Ulva (Chlorophyta) in intensive fishpond systems: mass culture and treatment of effluent. Journal Applied Phycology 2: 363 374. Von Harlem, O. (2006). Numerische Modellierung der Nährstoffdynamik ei ner integrierten Aquakultur Kreislaufanlage. M.Sc.thesis, Carl von Ossietzky Universität Oldenburg. Waller U., Bischoff A.A., Orellana J., Sander M., and Wecker B. (2003). An advanced technology for clear water aquaculture recirculation systems: Re sults from a pilot production of Sea bass and hints towards "Zero Dis charge". European Aquaculture Society Special Publications 33: 356 357. Waller U., Sander M., and Orellana J. (2005). A “low energy” commercial scale recirculation system for marine finfish. European Aquaculture Society Special Publications 35: 459 460. Watten B. J. (1992). Modelling the effects of sequential rearing of the poten tial production of controlled environment fish culture systems. Aquacultural engineering 11: 33 – 46. Wehner R. and Gehring W. (1995). Zoologie. Georg Thieme Verlag Stuttgard, New York. 861 pp.
88
Chapter 3 Cultivation of microalgae using a continuous photobioreactor system based on dissolved nutri ents of a marine recirculation system Kube N., Bischoff A.A., Wecker B., Waller U. (2006)
89 Chapter 3 Cultivation of microalgae using a continuous photobioreactor system based on dissolved nutrients of a marine recirculation system Kube, N., Bischoff A.A., Wecker B., Waller U. Abstract In this study a photobioreactor system for continuous microalgae cultivation was developed and tested in an integrated recirculation system with a water renewal rate of less than 1% system volume (MARE = Marine Artificial Recir culating Ecosystem). The photobioreactor system was equipped with a continuous water pre treatment unit and harvesting unit (foam fractionation) and tested for its ap plicability in aquaculture systems. Nannochloropsis was the target species and cultivated with a daily yield of 0.93 ± 0.5g dry weight m 2. C/N ratio, en ergy content and organic amount was determined for the harvest. Growth performance and fatty acid profiles of Nannochloropsis were determined ac cording to different nutrient concentrations at different irradiations (100, 250 and 400 mol s 1 m 2), and described by Michaelis Menten kinetics. This study showed that a continuous cultivation without automatic regula tion of light intensity and flow rates results in very unstable conditions re garding biomass concentration within the photobioreactors due to varying culture conditions. A semi continuous system might be more feasible.
90 Chapter 3 3.1 Introduction Microalgae play an important role in mariculture. They are used for rearing fish larvae, bivalves, shrimps and crabs (Brown et al., 1997; Moss, 1994) and for the nutrition of food organisms like rotifers and copepods (Støttrup and McEvoy, 2002; Pinto et al., 2001). Different species are used for this purpose (e.g. Nannochloropsis spec., Tetraselmis spec., Dunaniella spec., Isochrysis spec., (Brown et al. , 1989, Rocha et al. , 2003). Especially microalgae charac terized by an elevated content of polyunsaturated fatty acids are preferred (Volkman et al., 1992; Renaud et al. , 1994; 1999). Considerable populations of microalgae are cultivated not only for aquaculture purposes with standard cultivation methods (Becker, 1994), e.g. large pond cultures and highly effec tive photobioreactor systems (Borowitzka, 1999; Pulz, 2001). The usage of microalgae as a secondary integration step in marine recircula tion systems for removing dissolved nutrients from the system water is a new goal in aquaculture (Hussenot, 2003; Schneider 2005). High nutrient concentrations are present in the discharge water of recirculation systems: nitrate and phosphate are the two major nutrients which accumulate in a recirculation system. In comparison with conventional microbial biofilters, microalgae have several advantages: nitrate and phosphate are consumed simultaneously and used for incorporation into biomass (primary produc tion). Microalgae can also be regarded as a valuable “industrial” product with a variety of applications (food, cosmetics, human consumption, pharmacy etc.) (Wikfors and Ohno, 2001). Aim of this study was the development of a continuous cultivation system for marine microalgae based on dissolved nutrients from a marine recirculation system. A maximal flowback of purified water (without algae) needs to be achieved, because in modern marine recirculation systems water discharge is less than 1% per day (Waller et al. , 2003). New developments of systems for cultivation of microalgae in a secondary loop within a recirculation system are scarce because of the considerable technical effort for operation. Thus, these modules have been integrated into
91 Chapter 3 fishfarms only occasionally. First approaches have been performed, but were mainly focused on benthic diatoms, although commercial value of diatoms appears to be limited (Tandler, 2003; Hussenot, 2003). Microalgae with commercial value or required nutrient composition for aquaculture purposes are mainly floating algae. Due to their floating way of living, harvesting of the algae is a technical problem. Standard methods like filtration, sedimenta tion, flocculation or centrifugation require high technical/financial invest ments and are labour intensive (Becker, 1994). Therefore, integration of fil trating organisms into the recirculation system like bivalves are cheaper and more valuable alternatives for removing microalgae from the system water. (Pfeiffer and Rusch, 2000; Hussenot et al. , 1998; Hussenot, 2003). However, additional to the problems concerning the harvesting process of the algae, cultivation of free floating microalgae in fish effluent waters makes high demands compared to the cultivation using standard (sterile) media. Firstly, effluents from a fish tank are far from being sterile. Bacteria, fungi, viruses and particles may be transferred into the culture and may cause sig nificant difficulties during operation. Additionally, denitrification processes within a recirculation system can cause an inappropriate N:P ratio of the system water (e.g. 4:1; see Chapter 2). This leads to limitation of nutrients (N limitation), resulting in poor growth and sudden crashes of cultures (Bene mann 1992, own observations). In summary, the conceptional design of a photobioreactor system in a sec ondary loop of a marine recirculation system needs to fulfill several require ments. Technical aspects are: (1) efficient water pretreatment to reduce mi crobial contamination, (2) sufficient nutrient flow and (3) simple and efficient harvesting. However, for practical reasons the photobioreactor must be easy to operate and to clean. Also adaptability to different algal species should be possible. Nannochloropsis sp. was selected because of its profile of highly unsaturated fatty acids, its applicability as food and the wide range of temperatures and salinities tolerated by this species (Rocha et al., 2003). Nannnochloropsis sp.
92 Chapter 3 taxonomically belongs to the Eustigmatophyceae (Xanthophyceae, Chryso phyta). The flagellate cells (Fig. 1) are characterized by the possession of the xanthophylls heteroxanthine and vaucheriaxanthine and by a lack of chl b. Besides oil, chrysolaminarine and laminarine are accumulated as repository polymers.
Fig. 1: Nannochloropsis sp. (image source: www.innovations report.de ) 3.2 Material and Methods 3.2.1 Design of the continuous photobioreactor system The photobioreactor system used in this study consisted of three different units (Fig. 2): the "disinfection" unit, the microalgae cultivation unit and the harvesting unit. The system was operated as a bypass loop of the recircula tion system for continuous treatment of the process water from the recircu lation system.
The „disinfection“ unit (Fig. 1) was an 8L foam fractionator (Sander, Uetze Eltze, Germany) (1) combined with a water storage column. Water was per manently pumped (Eheim 1260) to the foam fractionator. Flow rate through the foam fractionator was manually adjusted to 32L/hour. The foam frac tionation was operated with a compressed air/ozone mixture, produced by an ozone generator (Sander, Ozonizer A500). Redoxpotential (ORP) was measured in the outlet of the foam fractionator by a sensor and regulated to reach values between 500 and 600mV using a control module KM2000 (both Sensortechnik Meinsberg GmbH, Ziegra Knobelsdorf, Germany).
93 Chapter 3
Waste air from the foam fractionation and the waste collector was passed to an ozone decomposer filled with activated carbon. The pretreated water was transferred to the storage column where water got aerated to release residual ozone and to stabilize pH. Remaining water in the water storage column was discharged via an overflow pipe into the primary recirculation system.
The microalgae cultivation unit consisted of three parallel photobioreactors (Fig. 1, B). Each photobioreactor consisted of an acrylic column with a di ameter of 20 cm and 1,50 m height, resulting in a total volume of 50 litres. The acrylic column was fixed with screws to a base plate of 50 x 40 cm. The top cover was designed to be removable. Y shape air spargers were attached in the lower part of the reactor, producing large air bubbles in order to agi tate/stir the culture. Four removable lamps (individually switched) with two fluorescent tubes each (daylight spectrum) were attached vertically around the bioreactor. To ensure maximal return of water back to the primary recirculation system, an efficient harvesting unit needed to be included. Foam fractionation was considered to be the most efficient method to harvest the suspended algae from the outflow water of the photobioreactors. A second foam fractionator identical to the first one (see above) was installed (Fig. 2 and 3a, (4)). The foam fractionator was also operated with an ozone/air mixture. Harvested algae were collected in a separate container (box “harvest”, Fig. 2). Purified water was led back to the primary recirculation system (sampling point C). A redox potential sensor was installed to control the redox potential in the out flow water.
94 Chapter 3
A 3 B
disinfection unit KM 2000 waste A 3 B microalgae cultivation unit 1 2 A ozone 3 B
recirculation system C 4 ozone harvest D harvesting unit
Fig. 2 Schematic drawing of the continuous photobioreactor system installed in a marine recircula tion system: (1) = foam fractionator for water pretreatment supplied with ozone; (2) = storage column with aeration to remove residual ozone, broken arrow indicates overflow for dispensable water; (3) = photobioreactors with attached light tubes; (4) = harvesting foam fractionator with ozone; grey circle with arrow = electrode for redox potential; white circle with arrow = electrode for pH; technical sym bols for pump (circle with triangle) and valves (double triangle), Arrows indicate flow of water. A,B,C,D = sampling points (Tab. 1).
95 Chapter 3
a 1 b 1 3 3
7
4 6
5 Fig. 3 Photobioreactor system: c (a) entire system with pretreatment foam frac tionator (1), two bioreactors (3) [one with light tubes, left), one without lights (right)] and har vesting foam fractionator (4); (b) disinfection foam 3 fractionator (1) with storage column with aeration (5) and overflow pipe (6), pump (7); (c) bottom sector of the bioreactors (3) with Y shape spargers and attached light sockets (without fluorescent tubes).
3.2.2 Functional principle of the photobioreactors The functional principle of the photobioreactor operation is analogous to a chemostat (Pirt, 1975, Fig. 4): medium is inserted into the reactor containing the culture at a constant flow rate (F). The total volume of the culture (V) remains constant by continuous removal following an overflow principle (Fig. 4). Thus, constant mixing processes are ensured in order to dispense the new media uniformly throughout the whole culture within a very short time.
96 Chapter 3
x = 0 x s s = s r mixing F F
medium culture
x s
Fig. 4 Simplified principle of a chemostat (Pirt, 1975): x = biomass, s = growth limiting substrate, s r = inflow concentration for growth limiting nutrient; F = flow rate; V = culture volume If a growing batch culture is changed into continuous cultivation modus, there are different scenarios to be considered as a result of the cultivation efforts: (1) the rate of washout exceeds the growth rate ( ) and biomass ac cumulation in the photobioreactor will decrease and growth limiting sub strate concentration will tend towards s r; (2) the initial rate of washout will balance the growth rate, so microalgae will grow at maximum rates ( m) and (3) the rate of washout is minor to the maximum growth rate, the biomass will increase within the photobioreactor. Therefore, growth of biomass can be described as follows for an infinitely small time interval dt : dx F × x = µ × x − (Equ. 1) dt V When the flow rate equalizes the specific growth in the photobioreactor, “steady state” is achieved. The steady state is a self regulating process: bio mass concentration and substrate concentration are acting following an os cillating relation: a decrease in biomass concentration will be associated with an increase in substrate concentration; this in turn will lead to an increase of the growth rate and thus, restore the steady state conditions. An increase in biomass concentration will have the reverse effects. Therefore, a certain time interval is required until this oscillation effect will tend to zero. So also
97 Chapter 3 biomass output possibility (1) and (3) will also result in a steady state at constant nutrient inflows conditions after a certain time. During the entire experimental period no real “steady state” can be expected to develop due to changing nutrient concentrations of the inflow process wa ter. Thus, measurements of the critical dilution rates (rates where diluting washout of the culture takes place) were used to determine growth rate at different concentrations of the recirculation water. Flow rates of recirculation water were adjusted to a value, at which optical density (OD 665 ) within the bioreactor remained constant for several hours. “Steady state” conditions were assumed and specific growth was determined by measuring the propor tion of microalgae removed from the photobioreactor within a certain time interval (1 hour):
ln N − ln N µ = t 0 (Equ. 2) t where N 0 = cell number at t 0 in the bioreactor, N t = cell number after t (calcu lated from cell numbers of bioreactor + cell numbers of outflow), t is time be tween two sampling points (1 hour).
Specific growth rates vary according to nutrient concentrations of the me dium. The relationship between specific growth rate and growth limiting nu trient can be described by a Michaelis Menten kinetic: µ ⋅ x µ = max (Equ. 3) + K s x where max = maximum growth rate, x = algae biomass, K s = half saturation constant. To calculate the nutrient concentrations (also called nutrient load) within the photobioreactor some additional aspects have to be taken into account. If a very small inflow volume is introduced into a larger volume (e.g. volume of the photobioreactor), dilution processes occur following an exponential equa 98 Chapter 3 tion: for example a dilution rate of 0.02 will not result in a 2% water ex change of the total photobioreactor volume. The percentage of exchanged wa ter (R) has to be calculated as a function of the flow rate (F) and the volume of a single photobioreactor (V) for a definite time interval (t):
−tF R = 1− e V (Equ. 4)
Considering this, the virtual effluent concentration (c out, virtual) describes the hypothetic effluent concentration at the end of the sampling period, assum ing that no algae are present. The function requires the measurement of the influent (c in ) and effluent (c out ) concentration at the beginning (t 1) and the end (t 2) of the sampling period. + + cin ,t cin ,t cout ,t cout ,t c = R ⋅ 1 2 + ()1− R ⋅ 1 2 (Equ. 5) out ,virtual 2 2
The virtual effluent concentration (c out, virtual ) was used to calculate the nutri ent uptake rate (r A) as the difference between the virtual and the measured effluent concentration (c out,t 2) per time interval (t). − (cout ,virtual cout ,t ) r = 2 (Equ. 6) A t
This study was focused on the determination of specific growth rates of the microalgae with regard to different nutrient concentrations of the photobio reactor in order to define suitable dilution rates for the operation of the photobioreactors. This was tested at different radiation intensities (100 mol s 1 m 2, 250 mol s 1 m 2, 400 mol s 1 m 2). The photobioreactor system was attached to a marine recirculation system for rearing sea bream ( Sparus au- rata) (see Chapter 5). It was operated during a 4 months experimental pe riod. Each reactor was working with a different irradiation and was illumi nated for 24 hours (Table 1).
99 Chapter 3
Tab. 1 Number of fluorescent tubes for illumination and measured average irra diation Photobioreactor Number of fluores radiation ( mol s 1 m 2) cent tubes 1 2 100 2 4 250 3 8 400
3.2.3 Algae Nannochloropsis was obtained from a sterile stock culture strain, available at IFM GEOMAR. Stock cultures served as backup and were cultured on f/2 medium (Guillard, 1975) applying sterile conditions. Working cultures of 5 l volume for bioreactor inoculation were already grown on pretreated recirculation water.
3.2.4 Culture conditions Water temperature of the recirculation system was kept at 20 ± 2.2°C and salinity at 23 ± 1.0 psu. Temperature in the bioreactors varied between 23 and 26°C due to different radiation intensities. pH values ranged from 7.5 to 8.5 and was not regulated automatically, but HCl was added when pH rose above values of 8.7. Compressed air was supplied to the photobioreactors in order to aerate the cultures. Bioreactors were started as batch cultures until the exponential growth phase of microalgae was observed. When the exponential growth phase was reached, the photobioreactors were transferred to the continuous cultivation modus. Flow rates varied between 0.36L h 1 to 1.5L h 1.
3.2.5 Sampling and analytical methods The protocols of daily measurements are summarized in Table 2. Depending on the production mode (batch, continuous), samples were taken at different sampling points (A,B,C,D, Fig. 2).
100 Chapter 3
Tab. 2 Daily measurements according to cultivation modus of the photobioreactor system included into a marine recirculation system. A,B,C,D are the sampling points (see Fig. 2). OD 665 = optical den sity of microalgae cultures at 665nm wavelength. A B C D Cultivation mode (inflow to biore (bioreactors) (outflow to (harvest) actors) MARE)
OD 665 pH Batch dissolved nutri ents
OD 665
OD 665 OD 665 pH pH pH pH dry matter Continuous dissolved nutri organic matter ents dissolved nutri dissolved nutri ents ents C/N energy Optical density was measured at 665nm with a Hach spectrometer (10 ml sample). In order to allow rapid determination of cell numbers a linear rela tionship between cell counts and optical density was established. Cell num bers were counted using a light microscope and a haemocytometer (Fuchs Rosenthal). Water samples of approx. 20ml were centrifuged at 5000 rpm for 10min. To remove algae and then frozen at 20°C. Dissolved nutrients (PO 4 P, NO 3 N,
NO 2 N, TAN) were analysed using an autoanalyzer AA3 (Bran Lübbe GmbH, Norderstedt, Germany). pH was measured with a WTW multi 340 pH meter.
Daily water volume in the harvest collection unit (Fig. 2, D) and cell numbers of harvested microalgae were recorded. 12 subsamples of microalgae harvest (10ml each) were centrifuged at 5000 rpm for 10min. The microalgae pellets formed by centrifugation were analysed as follows: content of dry matter was determined by drying the sample in a furnace at 60°C for 12 hours. Organic content was then measured by incineration of organic matter in a muffle furnace. C/N ratio was determined by gas chromatography (GC) in an ele ment analyser (EURO EA elemental analyser, Milano, Italy). Calorific value was measured by combustion in an IKA calorimeter C4000. Weighing was performed using a Sartorius A 210 P (max. 200g).
101 Chapter 3
Microalgae samples from the photobioreactors were filtered using 25mm glass fibre filters (Whatman GF/F). The filters were pre combusted for at least 12 hours at 550°C to remove traces of organic matter. POC and PON were measured applying gas chromatography. POP was converted to ortho phosphate by cooking with potassium peroxydisulphate. The orthophosphate was then measured photometrically. Fatty acids were extracted with dichloromethane: methanol (2:1, vol/vol) as described by Fink (Fink, 2006). PUFAs (polyunsaturated fatty acids) were quantified as fatty acid methyl esters (FAMEs) using a HP 5890 Series II GC (Agilent Technologies, Waldbronn, Germany) equipped with a DB 225 fused silica column (J&W Scientific, Folsom, USA) and a flame ionisation detector with heptadecanoic acid methyl ester and tricosanoic acid methyl ester as internal standards (Von Elert and Stampfl, 2000). Identification of the FAMEs was based on comparison of retention times to those of reference compounds. 3.3 Results 3.3.1 Feasibility of the photobioreactor system for algae cultivation a) pretreatment of recirculation water and harvesting process The photobioreactor system fulfilled the technical requirements concerning continuous water treatment and harvesting as well as space requirements. In general, the conceptional design of the photobioreactor system turned out to be feasible. It was possible to successfully cultivate Nannochloropsis in a continuous culture based on dissolved nutrients derived from the marine recirculation system. Water from the detritivorous culture tank (harbouring Nereis diversicolor ) was continuously pretreated with ozone in combination using foam frac tionation. This treatment process was analyzed for its effectiveness in purify ing highly loaded waste water from recirculation systems. As a result, disin fecting effect could be achieved due to the removal of particles and bacteria cells by the foam fractionation process. For details see Chapter 4.
102 Chapter 3
The harvesting process could be successfully established. Harvest of all cul tivated algae by the foam fractionation technique was achieved, which is rep resented by an average optical density at 665nm (OD665 ) of 0.007 ± 0.007 (< 1 Mio. cells ml 1). The daily amount of algae biomass yield averaged 2.49 ± 1.34g DW d 1. Daily yields are shown in Tab. 3. At the beginning of the ex periments the harvesting foam fractionator was operated without ozone. This turned out not to be effective enough, resulting in remaining algae in the outflow and an elevated harvesting volume (2.5 Litres) with relatively low cell densities. Addition of ozone improved the performance. At the end of the ex perimental period the harvest volume reached 0.184 ± 0.053 L d 1 at very high cell densities of 227.9 x 10 8 ± 6.7 x 10 8 cells ml 1 (Tab. 3). 99.68 ± 0.38 % of the water coming from the photobioreactors were transferred back to the main recirculation system. b) nutritional value of the cultivated microalgae The nutritional quality of the algae was increasing during the experimental period. The organic proportion increased from 40 50% to more than 80% of algae DW algae (Tab. 3), coinciding with an elevated energy content of the mi croalgae. Measured energy values of 21.07 ± 4.6 KJ d 1 DW algae indicate high nutritional quality of the microalgae. In initial samples a proportion of an unknown white substance could be observed, which was probably lime. This may explain the lower organic fraction, but not the high C:N ratio of the samples, because samples were pretreated with HCl in order to remove inor ganic carbon before gaschromatographical determination.
Fatty acid profiles were determined at the end of the experiments to get an other hint about the nutritional value of the cultivated microalgae (Tab. 4).
Total amount of fatty acids varied from 16.72 ng fatty acids (FA)/ g POC algae
(100 mol s 1 m 2) to 14.01 ng FA/ g POC algae (250 mol s 1 m 2). Values for full illumination was very low at 5.5 ng FA/ g POC algae . Highest values for Arachidonic acid (ARA) and Eicosapentaenoic acid (EPA) have been found in bioreactors at low irradiances at 1.12 ± 0.18 ng ARA/ g
103 Chapter 3
POC algae (100 mol s 1 m 2) and 1.16 ± 0.06 ng ARA/ g POC algae (250 mol s 1m 2). Values for Eicosapentaenoic acid (EPA) were 2.54 ± 0.44 ng EPA/ g
POC algae and 2.66 ± 0.14 ng EPA/ g POC algae , respectively. Values for irradia tion at 400 mol s 1 m 2 were found to be dramatically low at 0.22 ± 0.04 ng
ARA/ g POC algae and 0.50 ng EPA/ g POC algae for EPA. Relative amounts of highly unsaturated fatty acids compared to the total amount of fatty acids was high: relative amount of arachidonic acid (ARA, 20:4 n 4) of the total fatty acid content were at 6.7% (100 mol s 1 m 2), 8.27% (250 mol s 1 m 2) and 4.0% (400 mol s 1 m 2), respectively. Values for eicosapentaenoic acid (EPA, 20:5 n 5) were at 15.2% (100 mol s 1 m 2), 18.9% (250 mol s 1 m 2) and 9.1% of total fatty acid amount (400 mol s 1 m 2), respectively.
104
Tab. 3 Harvesting results by foam fractionation (total amount of all 3 photobioreactors with different irradiations at 100 mol s 1 m 2, 250 mol s 1 m 2 and 400 mol s 1 m 2, respectively). Flow rate per day = total flow through all three bioreactors, foam volume = harvested volume in the har vesting vessel, cell density, pH, yield (DW d 1) in the harvest. Organic fraction, energy content and C/N ratio of the cultivated algae. Date Flow rate Foam vol Cell density pH Yield DW Org. frac Energy N C C:N ratio per day [L] ume [L] [Mio ml 1 Harvest (g d 1) tion (KJ g 1) [mg g 1 DW] [mg g 1 DW] harvest] 15.10.05 108.0 1.0 129.1 7.09 4.44 55.8 16.05 36.06 317.4 8.8 17.10.05 108.0 0.4 164.6 7.46 2.58 57.4 14.21 38.84 297.8 7.7 19.10.05 108.0 0.12 126.1 n.d. 0.62 50.2 13.85 29.62 263.8 8.9 20.10.05 108.0 0.58 133.8 7.09 3.01 53.6 13.22 24.90 213.1 8.6 27.10.05 73.4 2.5 68.8 7.15 n.d. 39.9 13.54 15.36 143.8 9.4 28.10.05 73.4 1.5 392.2 7.06 n.d. 53.9 9.16 31.16 298.2 9.6 29.10.05 73.4 0.98 115.6 7.3 2.69 83.4 14.27 45.01 386.2 8.6 31.10.05 73.4 1.75 65.3 7.42 3.07 78.9 23.53 52.46 458.1 8.7 1.11.05 73.4 0.35 207.4 7.06 1.69 87.9 24.63 58.8 499.3 8.5
105 2.11.05 73.4 0.4 214.2 7.15 3.44 82.9 22.88 57.2 473.3 8.3 16.11.05 260.8 0.8 25.0 7.03 82.5 21.39 49.8 408.2 8.2 18.11.05 108.0 0.4 31.80 n.d. 4.58 78.5 21.49 55.1 417.3 7.6 20.11.05 86.4 0.49 46.6 6.93 7.41 78.5 21.66 49.5 416.3 8.4 21.11.05 86.4 0.18 32.7 6.81 3.02 73.4 21.23 36.5 307.2 8.4 24.11.05 86.4 0.2 35.3 6.81 2.74 82.1 23.46 45.3 341.9 7.6 25.11.05 86.4 0.2 29.6 6.38 1.86 85.1 23.88 63.9 480.0 7.5 26.11.05 86.4 0.3 27.7 6.55 3.16 80.6 21.93 52.0 381.5 7.3 27.11.05 86.4 0.16 56.6 6.87 2.62 85.9 23.80 80.3 593.6 7.4 28.11.05 86.4 0.05 75.3 5.88 1.48 91.9 25.98 50.7 407.5 8.0 10.1.06 43.2 0.16 355.0 6.85 2.38 11.1.06 43.2 0.15 466.1 n.d. 1.45 85.1 23.61 53.8 463.0 8.6 16.1.06 64.8 0.18 704.1 6.81 1.58 84.0 22.82 62.8 506.6 8.1 17.1.06 43.2 0.22 674.5 6.94 1.88 84.8 24.19 65.6 472.8 7.2 18.1.06 60.5 0.19 667.9 7.43 1.56 85.1 23.78 60.6 448.6 7.4
Date Foam vol Cell density PH Yield DW Org. frac Energy N C C:N ratio ume [L] [Mio ml 1 Harvest (g d 1) tion (KJ g 1) [mg g 1 DW] [mg g 1 DW] harvest] 19.1.06 64.8 0.2 901.4 7.35 2.31 84.7 24.48 56.6 432.6 7.6 20.1.06 64.8 0.1 954.8 6.96 1.15 86.9 25.12 72.9 538.6 7.4 22.1.06 64.8 0.18 849.6 5.48 1.77 87.8 24.96 75.1 534.7 7.1 23.1.06 64.8 0.11 772.3 7.92 0.98 86.7 24.8 65.3 493.5 7.6 24.1.06 64.8 0.31 683.5 5.66 2.43 85.7 24.12 67.8 482.7 7.1 25.1.06 64.8 0.18 760.0 n.d. 2.10 84.9 22.97 64.3 468.6 7.3 Tab. 4 Fatty acid profiles of Nannochloropsis spec. from batch cultures at stationary phase at different irradiances (100 mol s 1 m 2, 250 mol s 1 m 2, 400 mol s 1 m 2) Fatty acid: posi nomenclature 100 mol s 1m 2 250 mol s 1 m 2 400 mol s 1 m tion of double 2 bonds average ± SD average ± SD average ± SD ng FA/ g PO ng FA/ g PO ng FA/ g PO 106 Calgae Calgae Calgae 14:0 Myristic acid 0.23 ± 0.32 0.15 ± 0.34 16:0 Palmitic acid 5.75 ± 0.79 5.12 ± 0.33 2.01 ± 0.46 16:1 Palmitoleic acid 3.44 ± 0.34 2.76 ± 0.17 0.68 ± 0.16 18:1 (9)/(6) Oleic acid 2.96 ± 0.40 1.71 ± 0.09 0.93 ± 0.20 18:1 (11) Vaccenic acid 0.14 ± 0.08 0.10 ± 0.01 0.37 ± 0.07 18:3 (6,9,12) y Linolenic acid 0.03 ± 0.04 0.09 ± 0.00 18:3 (9,12,15) a Linolenic acid 0.51 ± 0.06 0.26 ± 0.01 0.79 ± 0.37
20:4 (5,8,11,1) Arachidonic acid 1.12 ± 0.18 1.16 ± 0.06 0.22 ± 0.04 (ARA) 20:5 (5,8,11,14,17) Eicosapentaenoic 2.54 ± 0.44 2.66 ± 0.14 0.50 ± 0.10 acid (EPA) Total 16.72 14.01 5.5
Chapter 3
3.3.2 Specific growth rates and nutrient uptake rates of Nannochlorop- sis at different light intensities The duty of a continuous photobioreactor system in a marine recirculation system is to maintain cultivation rates at constant rates with maximum yield rates. Therefore adequate flow rate according to the available nutrient con centration and hence specific growth rate needs to be adjusted: if the flow rate is higher than the specific growth rate of the microalgae, washout of biomass will set in and may result in a low performance of the photobioreac tor. Flow rates where washout happens are called “critical dilution rates” and were determined in this study at various nutrient concentrations within the photobioreactor by adjusting different flow rates during the four month ex perimental period at different irradiations (100 mol s 1m 2, 250 mol s 1m 2 and 400 mol s 1m 2) in order to obtain information about maximum specific growth rates within this type of photobioreactor.
Data are presented exemplary from the photobioreactor with the lowest irra diation (100 mol s 1m 2, Fig. 4, Tab. 3). Batch culture (black dots) was oper ated until a sufficient optical density was reached. During period A (Fig. 5) flow rate of 1.5 Litre per hour was adjusted at day 8 and resulted in a rapid decrease of biomass within 4 days. Therefore, continuous mode was changed to batch culture to increase cell numbers. At day 21 flow rate of 1 Litre/hour was adjusted (period B). For the initial cultivation days biomass remained constant, indicating that flow rate equalized specific growth rate of microal gae. From day 23 onwards a decrease of specific growth rates was observed. This is due to a slow but constant (until day 30) washout of algal biomass. Another continuous cultivation period was started at day 42 (period C) with a flow rate of 3.6 Litres per hour resulting in a rapid biomass decrease within a short time period. Flow rate was adjusted to 1.2 Litres/hour (period D). At this flow rate biomass values in the photobioreactor remained con stant. Cultivation was stopped at day 58 due to a system failure. During that period water parameters (dissolved nutrients, pH etc., Tab. 4) remained con stant.
107 Chapter 3
In periods E,F,G flow rates needed to be adjusted to lower values, because of limited algae growth, because nitrate was limited in the recirculation system (see Chapter 5) (Tab. 4, period E,F,G). This resulted in lower growth rates of Nannochloropsis sp.
14 100 µmol s -1 m-2 batch 12 continuous
10
-1 8
6
4 cells Mio ml cells Mio
2
0 A B C D EFG
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140
days Fig. 5 Cell densities of Nannochloropsis per ml in the photobioreactor illuminated with 2 fluorescent tubes (100 mol s 1m 2). Batch cultures (black dots) were used to reach sufficient densities, before operation mode was changed to continuous production (white dots). Letters indicate adjusted flow rates within the periods with continuous production: A = 1,5L/hour (dilution rate 0.03); B = 1L/hour (dilution rate 0.02); C = 3.6L/hour (dilution rate 0.08); D = 1.2L/hour (dilution rate = 0.027); E = 0.6L/hour (dilution rate 0.027); F = 0.9L/hour (dilution rate = 0.02); G = 0.36L/hour (dilution rate = 0.014) (for further details, see also Tab. 4).
108 Chapter 3
Tab. 4 Water parameters during continuous production of Nannochloropsis spec. in the photobioreactor with low illumination (100 mol s 1m 2) during the 4 month experimental time. A,B,C,D,E,F,G are indicating different time intervals with different adjusted flow rates. D = dilution rate/h Flow inflow concentration outflow concentration Period rate D pH to photobioreactor from photobioreactor L/h
PO 4 P NO 3 N TAN NO 2 N PO 4 P NO 3 N TAN NO 2 N
A 1.5 0.03 7.44 25.2 60.5 n.d. n.d. 19.5 56.2 n.d. n.d. ±0.05 ± 2.2 ± 2.0 ± 19.5 ±3.0 B 1.0 0.02 8.01 30.6 80.7 0.93 0.15 20.5 71.4 0.17 0.19 ±0.21 ± 3.3 ± 8.4 ± 0.58 ± 0.08 ± 2.6 ± 8.4 ±0.08 ± 0.12 C 3.6 0.08 7.8 32.2 82.4 0.3 0.5 19.9 72.6 0.1 1.4 ± 0.1 ± 2.3 ±1.8 ± 0.2 ± 0.4 ± 2.4 ± 5.4 ± 0.03 ± 0.6 D 1.2 0.027 7.88 37.3 92.1 0.14 0.07 22.1 87.1 0.08 0.18 ± 0.2 ±1.28 ± 4.7 ± 0.06 ± 0.06 ± 4.3 ± 1.6 ± 0.05 ± 0.21 E 0.6 0.022 8.4 28.9 12.2 2.6 0.01 11.5 36.2 0.1 0.4 ± 0.4 ± 4.8 ± 1.6 ± 0.5 ± 0.01 ± 3.0 ± 3.0 ± 0.1 ± 0.1 F 0.9 0.02 7.51 33.0 8.9 1.7 0.04 14.6 5.5 0.11 0.06 ± 1.4 ± 6.3 ± 2.0 ± 0.97 ± 0.05 ± 7.2 ± 2.91 ± 0.11 ± 0.03 G 0.36 0.014 8.12 41.2 9.5 3.7 0.11 25.9 0.00 0.06 0.06 ± 0.1 ± 3.4 ± 0.9 ± 0.4 ± 0.07 ± 1.7 ± 0.2 ± 0.04 ± 0.03 Photobioreactors with irradiations of 250 mol s 1m 2 and 400 mol s 1m 2 showed similar performances, although washout rates differed. For determi nation of specific growth rates, time intervals with constant cell densities within the photobioreactor at a certain flow rate were chosen. Specific growth rate ( ) was determined for 1 hour intervals according to equation 2. These values were plotted against the nutrient concentrations within the photobioreactor according to equations 4 and 5. Fig. 6 shows Michaelis Menten kinetics determined for each photobioreactor for nitrogen (sum of N0 3 N, TAN, NO 2 N). Kinetics for all three treatments turned out to be rather similar, maximum specific growth rates for all three radiation intensities were approx. 0.024h 1. Best growth and nutrient re moval rates have been determined for the photobioreactor operated at 400 mol s 1 m 2 irradiation for short time periods. However, photobioreactor at high illumination was very unstable, wash out of biomass occurred very
109 Chapter 3 often. Best long term performance was observed for photobioreactor with 250 mol s 1 m 2 irradiation.
0,035 0,035 a 100 µmol s -1 m-2 b 250µmol s -1 m -2 0,030 0,030
0,025 0,025
0,020 0,020
0,015 0,015
specific growth 0,010 0,010 specific growth rate
0,005 0,005
0,000 0,000 0 20 40 60 80 100 0 20 40 60 80 100 -1 -1 N concentration in the photobioreactor in mg L N concentration in the photobioreactor mg L
0,035 c 400 µmol s -1 m-2 0,030
0,025
0,020
0,015
0,010 Specific growth rate 0,005
0,000 0 20 40 60 80 100
N concentration in photobioreactor mg L -1 Fig. 6 Growth of Nannochloropsis spec. in continuous cultures of a photobioreactor system included in a marine recirculation system at different nutrient concentrations in the photobioreactor at different irra diations (100 mol s 1 m 2, 250 mol s 1 m 2, 400 mol s 1 m 2). Ks values for all three kinetics indicate that nutrient limitation is normally not occurring within a recirculation system, as system water concentrations of nitrogen and phosphorus are normally above limiting values. However, very low nitrogen concentrations may occur, if an enhanced denitrification occurs within a recirculation system, as happened in the experimental trial within this study was accomplished (MARE II, see Chapter 5). Nitrogen con centrations tended toward zero, resulting in a lower growth performance of the microalgae.
110
Tab. 5 Growth performance of Nannochloropsis spec. in continuous cultures in photobioreactors included in a marine recirculation system at different irradian ces (100 mol s 1 m 2, 250 mol s 1 m 2, 400 mol s 1m 2) = specific growth rate, max = maximum growth rate, K s = half saturation constant (Michaelis Menten kinetic), SE = Standard Error; SD = Standard Deviation a = mean of experimental data from continuous culture (n= 35), b = parameters from regression analysis (Sigma Plot) irradiance 100 Wm 2 250 Wm 2 400 Wm 2 1 1 1 1 1 1 h max h Ks h max h Ks h max h Ks ± SD ± SE ± SE ± SD ± SE ± SE ± SD ± SE ± SE continuous 0.020 0.024 0.41 0.020 ± 0.024 ± 2.01 0.021 ± 0.024 ± 0.53 ± ± 0.008 a ± 0.001 b ± 0.17 b 0.009 a 0.001 b ± 0.59 b 0.008 a 0.001 b 0.15 b
111
Tab. 6 Calculation of nutrient uptake (r) per hour per litre of Nannochloropsis spec. in continuous cultures in photobioreactors included in a marine recirculation system at different irradiances (100 mol s 1m 2, 250mmol s 1m 2, 400 mol s 1m 2). r = uptake rate per hour, r max = maximum growth rate; SD = Standard Devia tion a = mean of experimental data from continuous culture (n = 50); b = maximum observed value (n=1); irradiance 100 Wm 2 250 Wm 2 400 Wm 2
r r max r rmax r rmax r rmax r rmax r rmax
PO 4 P PO 4 P NO 3 N NO 3 N PO 4 P PO 4 P NO 3 N NO 3 N PO 4 P PO 4 P NO 3 N NO 3 N mg L 1h 1 mg L 1 h 1 mg L 1 h 1 mg L 1 h 1 mg L 1 h 1 mg L 1 h 1 mg L 1 h 1 mg L 1 h 1 mg L 1 h 1 mg L 1 h 1 mg L 1 h 1 mg L 1 h 1 ± SD ± SD ± SD ± SD ± SD ± SD continuous 0.190 0.41 b 0.129 0.29 b 0.219 0.52 b 0.135 0.48 b 0.294 0.59 b 0.186 0.51 b ± 0.11 a ± 0.09 a ± 0.16 a ± 0.14 a ± 0.15 a ± 0.15 a
Chapter 3 3.4 Discussion 3.4.1 Applicability of the photobioreactor design In this study, Nannochloropsis sp. was cultivated for the first time based on dissolved nutrients from a marine recirculation system. The conceptional design of the photobioreactor system was feasible concerning the pretreat ment of recirculation water for microalgae cultivation and the harvesting method of microalgae. For both processes, water pretreatment and harvest ing, the foam fractionation technique was applied. Foam fractionation was also applied in previous studies for harvesting of algal cells (Csordas and Wang, 2004) and turned out to be a very effective harvesting method. Very dense algal suspensions were harvested at minimum volumes, the outflow water from the foam fractionators was clear. Thus, foam fractionation proved to be a very effective method for microalgae removal. However, there are other techniques to be considered within an integrated system. Filtrating or ganisms like bivalves can use microalgae directly, and might be an addi tional trophic level (Hussenot et al., 1998, Pfeiffer and Rusch, 2000). There are several types of bioreactors, most of them reported to have very high yield rates. These reactors are often operated at optimum light intensi ties (Richmond et al. , 2003; Pulz, 2001). Especially Nannochloropsis was effi ciently cultivated in flat plate glass reactors (Cheng Wu et al. , 2001). Re ported cell densities for polyethylene bags or transparent fibreglass columns vary between 25 150 x 10 6 cells per ml 1 (0.1 to 0.6g ash free dry mass L 1), in flat plate reactors cell density can reach 5 to 6 x 10 8 cells ml 1 (12g dry cell mass m 2 day 1 or 650mg EPA m 2 day 1 (Cheng Wu et al. , 2001). These values could not be achieved during this study, in average 0.94g dry cell mass per m 2 and day could be harvested. This is considered to be mainly caused by the larger diameter of the photobioreactor columns, resulting in a comparably long distance for the light to reach all algal cells. Although the culture has been agitated by air bubbles, self shading of the algae induced photolimitation. In high tech bioreactors the equal supply of all algal cells with light has to be guaranteed to ensure maximum growth rates of the mi
112 Chapter 3 croalgae (e.g. 1cm in flat plate reactors, Richmond et al. , 2003). However, cultivation in long tubular systems with a low diameter has also already been reported to show acceptable growth rates (Pulz, 1994; Borowitzka, 1996). These systems have high spatial requirements, which are often lim ited in aquaculture systems. In this study, it was not possible to install such photobioreactor systems due to spatial limitations of the laboratory. A comparably large diameter of 20cm of the photobioreactor column was chosen in order to increase the culture volume. Hence a total volume of 50 Litres could be placed in a room providing an area of only 0.16m². The re duction of the optical path (equal supply with light for all algal cells) was compensated by agitating of the culture. Air bubbles were constantly di rected through the tank to ensure an equal light supply of all algal cells. This “bubble column system” has also been applied during cultivation of microal gae and was considered to be suitable (Eriksen et al. , 1998). However, a strong agitation can cause cell damage and hence the intensity of the bub bling has to be carefully adjusted. The culture has to be monitored cautiously. During cell division Nan- nochloropsis sp. new cell walls are generated and the “old” multilayered par ent cell walls are released. The release of this parent cell wall coincides with the release of autoinhibitory substances (Rodolfi et al. , 2003). When cell den sities reach a critical value, further growth can be inhibited by autoinhibiting substances released by the algae. Thus, both cell wall remains and autoin hibitors may negatively affect culture growth. Additionally, biofouling at the walls of the photobioreactor could be observed and may be another factor influencing continuous growth once maximum growth rates were reached. Because of this biofouling process, photobioreactors had to be cleaned every 2 to 3 weeks. 3.4.2 Nutritional value of microalgae
113 Chapter 3
The harvested algae had a high energy content of up to 20kJ g 1 DW. This comparably high nutritional value makes the algae a suitable feed for other feed organisms like Brachionus or Copepods (Støttrup & Norsker, 1997). Nannochloropsis species are known for the high content of highly unsatu rated fatty acids (HUFA): 20:4n 6 have a proportion of 1.9 3.9% (by weight of total fatty acids) and 20:5n 3 contents (12.1% 17.8% as weight percentage of total fatty acids) at a total lipid content of 10.3 16.1% (dry weight, Mourente et al., 1990). Similar values have been observed in the cultivated algae dur ing this study: the relative proportion of arachidonic acid (ARA, 20:4 n 4) of total fatty acid amount were at 6.7% (100 mol s 1 m 2), 8.27% (250 mol s 1 m 2) and 4.0% of total fatty acid amount (400 mol s 1 m 2), respectively. Simi lar values were determined for eicosapentaenoic acid (EPA, 20:5 n 5) at 15.2% of total fatty acid amount (100 mol s 1 m 2), 18.9% (250 mol s 1 m 2) and 9.1% of total fatty acid amount (400 mol s 1 m 2), respectively. The nu tritional value of the cultivated microalgae was of special interest because of the applicability for successful fish larvae rearing (Støttrup and McEvoy, 2002). The decreasing content of HUFA with increasing light intensities was already reported before (Fabregas et al. , 2003). Illumination exceeding light satura tion (>220 mol s 1 m 2) induces a dramatically decrease in the content of highly unsaturated fatty acids in the microalgal cells. Temperature also is reported to influence fatty acid synthesis in Nannochloropsis : with increasing temperatures up to 25°C an increasing trend in the content of HUFAs can be shown, but then a rapid decrease occurs when temperatures of 25°C are ex ceeded (James et al. , 1989). Also salinity has an influence on the content of HUFAs. Optimum salinities for Nannochloropsis oculata were reported to be 20 30ppt (Renaud and Parry, 1994). These salinities were therefore chosen for cultivation of Nannochloropsis sp. in the photobioreactors during this study.
114 Chapter 3
3.4.3 Growth performance of Nannochloropsis spec. in continuous cul- tures The specific growth rates determined from continuous cultures can only be regarded as approximate values characterizing a small time interval of sev eral hours. The maximum specific growth rate of 0.025 h 1 can be reached at all irradiations, if the other cultivation conditions are suitable (e.g. nutrient supply, cell density, light intensity). But for long term cultivation periods, these values might be overestimated, because density in the photobioreac tors was always decreasing after a certain time probably partly due to the release of autoinhibitory substances. Another explanation is the absence of an automatic regulation of flow rate and illumination: it was not possible to reach a real “steady state” in the photobioreactors due to varying nutrient concentrations of the inflow water coming from the recirculation system. Generally, biomass and nutrient con centrations tend towards the steady state conditions with an oscillating process (Pirt, 1975). However, changes in cultivation conditions, especially nutrient concentrations, may disturb this process. Considering constant cul tivation of microalgae at high biomass levels, a disturbance may result in an outwash of biomass. Steady state conditions are unlikely to develop in a re circulation system, where several conditions can change within a very short time (e.g. diurnal variation of ammonia, increasing temperatures (optimum 15 20°C, James et al. , 1989) etc.). These condition changes are followed by a decreased growth rate due to a lag phase needed by the algae to adapt to the new condition. This lag phase leads to an wash out of biomass in a continu ous culture with a constant flow, if flow rate is constant in turn resulting in a reduced cell density in the photobioreactor. This cascade effect will con tinue: if the irradiation is not suitable for the “new” reduced cell density and is not adjusted, growth rate will continuously decrease (due to light limita tion or photoinhibition, Richmond et al. 2003). Biomass will then be washed out even more rapidly. This process is considered to be the reason for the unstable performance of the photobioreactor at high irradiations (400 mol) during long term experiments.
115 Chapter 3
Algae growth in the photobioreactor at the lowest irradiation (100 mol s 1 m 2) was affected by light limitation, but this photobioreactor operated rather stable in long term experiments compared to the bioreactors being treated with elevated radiation intensities. At this radiation intensity the algal growth rate may increase at a reduced velocity due to the lower il lumination. The photobioreactor at 250 mol s 1 m 2 radiation intensity showed best long term performance during the experiment, because light intensities fitted best to the varying cultivation conditions (temperature, nu trient availability), although biomass wash out also occurred. These results fit to a laboratory study where light saturation for Nan- nochloropsis was achieved at 220 mol s 1 m 2 with no significant increase in steady state cell density or dry weight productivity at higher irradiations up to 480 mol s 1 m 2 (Fabregas et al. , 2004). Similar results have been found in this study, although nutrient uptake rates can reach much higher values at irradiation of 400 mol s 1m 2.
In this study, growth of Nannochloropsis sp. was limited by nitrogen and could be described by a Michaelis Menten kinetic in relation to total nitrogen concentration, because in the photobioreactors different chemical forms of nitrogen (nitrate, ammonium and nitrite) were supplied to Nannochloropsis. It could not be proved that Nannochloropsis prefers one of these form, al though according to the data hints are given for nitrate preference. In na ture, summer and autumn phytoplankton blooms exhibits typical Michelis
Menten uptake kinetics for NO 3 , NH 4+ and urea, although also linear uptake rates may occur with increasing nitrogen concentrations due to eutrophica tion (Horner Rosser, 2004). Similar to the conditions in the presented photo bioreactor, growth rates and nutrient uptake rates in natural systems are not in steady state but will fluctuate according to large and small scale spa tial and temporal variations in nutrient availability. Certain species (e.g. Clamydomonas ) are able to maximise growth through short term rapid utili sation. This is considered to be a mechanism for competitive success in a mixed population (Rhee et al., 1988). Nannochloropsis was found to grow at
116 Chapter 3 similar growth rates at nitrate, ammonia and urea (Lourenco et al., 2002), indicating an acclimation of microalgae to the respective N sources (Levas seur et al. , 1993). In general, ammonium is the preferred chemical form of nitrogen and is readily taken up and assimilated by phytoplankton (Collos and Slawyk, 1981). In contrast to nitrate, microalgae do not need to reduce ammonia prior to assimilation (Syrett, 1981). However, some studies indicate that microalgae growth can be negatively influenced at high ammonia con centrations due to toxification effects (e.g. in media used in aquaculture or nutrient loaded effluents). Lourenco et al. (2002) showed that growth of Nan- nochloropsis oculata on urea resulted in a lower final biomass yield. How ever, supply of ammonia and urea was very low in the photobioreactor sys tem, so toxification effects are considered to be unrealistic.
It has been shown in few studies that different nitrogen sources may result in effects on growth and biochemical composition (content of protein, lipids and carbohydrates, fatty acid profiles) (Lourenco et al., 2002; Fidalgo et al., 1995). High nitrogen concentrations in the culture medium stimulates the accumulation of protein by microalgae (Fabregas et al. , 1989; Fidalgo et al. , 1995). This might be one reason of the high energy content of the microal gae.
Results from this study showed that uptake rates for phosphorus were higher than for nitrate. This was unexpected with regards to the Redfield ra tio. However, it has been shown that in nutrient rich water phosphorus can be rapidly taken up, resulting in a depletion of external phosphorus concen tration in the medium. Phosphorus can be stored inside the cells (Passarge et al. , 2006). Also, high intracellular concentrations of inorganic nitrogen can be built up if the nitrogen is available in large amounts in the medium (Lourenqo et al., 1998), but nitrogen was limited at the end of the experimen tal period (see Chapter 5).
117 Chapter 3
3.4.4 Filter efficiency of microalgae photobioreactors In contrast to the macroalgae filter, microalgae are able to remove nitrate from the water due to longer retention times of nitrate within the culture. According to the specific growth dilution rates in the photobioreactors may not exceed 0.025 h 1. For the presented photobioreactor system with a total volume of 150 Litres the flow rate per hour was appr. 3.75 Litres. This flow rate is far too low to be an effective biofilter for the removal of toxic ammonia and nitrite. However, daily removal rates of nitrate and phosphate have been determined at 0.87g PO 4 P d 1 and 0.57g NO 3 N d 1, respectively, supporting the importance to control nitrate and phosphate concentrations within a re circulation system. Thus, simultaneous integration of microalgae photobio reactors and macroalgae filters into a recirculation system is recommended to achieve an optimal removal of all dissolved nutrients derived from fish cultivation (see also MARE II experiment, chapter 5). 3.5 Conclusion The continuous culture mode turned out not to be suitable for a long term cultivation of microalgae within a recirculation system, if flow rate and light intensities are not automatically adjusted. It is technically possible to install an online flow through photometer and a computer controlled light regula tion into a recirculation system. However, these installations are expensive and may exceed investment costs for aquaculture systems. Better results may be achieved with semicontinuous microalgae cultures, where a certain percentage of the total volume is replaced every day.
3.6 Acknowledgements This study was funded by the EU (Interreg IIIa). We thank Line Christensen and Jens Jorgen Lonsman Iversen from the Syddansk Universitet Odense for supporting with knowledge about bubble column photobioreactors. We thank Kerstin Nachtigall and Peter Fritsche (IFM GEOMAR) for analysis of POC, PON and POP and Patrick Fink (Max Planck Institute Plön) for deter mining fatty acid profiles.
118 Chapter 3 3.7 References Becker E.W. (ed) (1994). Microalgae: biotechnology and microbiology. Cam bridge University Press. 287pp. Benemann J.R. (1992). Microalgae aquaculture feeds. Journal of Applied Phycology 4: 233 245. Borowitzka M.A. (1996). Closed algal photobioreactors: design considerations for large scale systems. Journal of Marine Biotechnology 4, 185 191. Borowitzka M.A. (1999). Commercial production of microalgae: ponds, tanks, tubes and fermenters. Journal of Biotechnology 70: 313 321. Brown M.R., Jeffrey S.W., and Garland C.D. (1989). Nutritional aspects of microalgae used in mariculture: a literature review. CSIRO Marine Laborato ries Report 205. 44pp. Brown M.R., Jeffrey S.W., Volkman J.K., and Dunstan G.A. (1997). Nutri tional properties of microalgae for mariculture. Aquaculture 151: 315 331. Cheng Wu Z., Zmora O., Kopel R., Richmond A. (2001). An industrial size flat plate glass reactor for mass production of Nannochloropsis sp. (Eustig matophyceae). Aquaculture 195: 35 49. Csordas A. and Wang J.K. (2004). An integrated photobioreactor and foam fractionation unit for the growth and harvest of Chaetoceros spp. in open systems. Aquacultural Engineering 30: 15 30. Collos Y. and Slawyk G. (1980). Uptake and assimilation by marine phyto plankton. In: Falkowski P.G. Primary productivity in the sea. Plenum Press, New York. Pp. 195 211. Eriksen N.T., Poulsen B.R., and Lønsmann Iversen J.J. (1998). Dual sparging laboratory scale photobioreactor for continuous production of mi croalgae. Journal of Applied Phycology 10: 377 328. Fabregas J., Abalde J., and Herrero C. (1989). Biochemical composition and growth of the marine microalgae Dunaliella tertiolecta (Butcher) with different ammonium nitrogen concentrations as chloride, sulphate, nitrate and car bonate. Aquaculture 83: 289 304. Fabregas J., Maseda A., Domínguez A, and Otero A. (2004). The cell compo sition of Nannochloropsis sp. Changes under different irradiances in semi continuous culture. World Journal of Microbiology and Biotechnology 20: 31 35.
119 Chapter 3
Fidalgo J.P., Cid A., Abalde J., and Herrero C. (1995). Culture of the marine diatom Phaeodactylum tricornutum with different nitrogen sources: growth, nutrient conversion and biochemical composition. Cahiers de Biologie Ma rine 36: 165 173. Fink, P. (2006). Food quality and food choice in freshwater gastropods: Field and laboratory investigations on a key component of littoral food webs. Ber lin, Germany, Logos Verlag, Berlin. Guillard, R.R.L. (1975). Culture of phytoplankton for feeding marine inverte brates. Pp 26 60. In: Smith, W.L. and Chanley, M.H. (eds.) Culture of Marine Invertebrate Animals. Plenum Press, New York, USA. Horner Rosser J. (2004). Phytoplankton Ecology in the Upper Swan River Estuary, Western Australia: with Special Reference to Nitrogen uptake and Microheterotroph Grazing. PhD thesis. Curtin University of Technology. 265pp. Hussenot J.M.E., Lefebvre S., and Brossard N. (1998). Open air treatment of wastewater from land based marine fish farms in extensive and intensive systems: current technology and future perspectives. Aquating Living Re sources 11: 297 304. Hussenot J.M.E. (2003). Emerging effluent management strategies in marine fish culture farms located in European coastal wetlands. Aquaculture 226: 113 128. James C.M., Al Hinty S., and Salman A.E. (1989). Growth and ω3 fatty acid and amino acid composition of microalgae under different temperature re gimes. Aquaculture 77:337 351. Levasseur M., Thompson P.A., and Harrison P.J. (1993). Physiological accli mation of marine phytoplankton to different nitrogen sources. Journal of Phycology 29: 587 595. Lourence S.O., Barbarino E., Lanfer Marquez U.M., and Adaire E. (1998). Distribution of intracellular nitrogen in marine microalgae: basis for the cal culation of specific nitrogen to protein conversion factors. Journal of Phycol ogy 34: 798 811. Lourenco S.O., Barbarino E., Mancini Filho J., Schinke K.P., and Aidar E. (2002). Effects of different nitrogen sources on the growth and biochemical profile of 10 marine microalgae in batch culture: An evaluation for aquacul ture. Phycologica 41: 158 168. Moss S.M. (1994). Growth rates, nucleid acid concentrations, and RNA/DNA ratios of juvenile white shrimp, Penaeus vannamei Boone, fed different algal diets. Journal of Experimental Marine Biology and Ecology 182: 193 204.
120 Chapter 3
Mourente G., Lubián L.M., and Odriozola J.M. (1990). Total fatty acid com position as a taxonomic index of some marine microalgae used as food in marine aquaculture. Hydrobiologica 203:147 154. Passarge J., Hol S., Escher M., and Huisman J. (2006). Competition fo nutri ents and light: stable coexistence, alternative stable states, or competitive exclusion? Ecological Monographs 76: 57 72. Pfeiffer T.J. and Rusch K.A. (2000). An integrated system for microalgal and nursery seed clam culture. Aquacultural Engineering 24: 15 31. Pinto C.S.C., Souza Santos L., and Santos P.J.P. (2001). Development and population dynamics of Tisbe biminiensis (Copepoda: Harpacticoida) reared on different diets. Aquaculture 198: 253 267. Pirt S.J. (1975). Principles of microbe and cell cultivations. Blackwell Scien tific publications. 284pp. Pulz O. (1994). Open air and semi closed cultivation systems for the mass cultivation of microalgae. In: Phang S.M., Lee K., Borowitzka M.A., Whitton B. (eds). Algal biotechnology in the Asia Pacific Region. Institute of Advanced Studies, University of Malaya, Kuala Lumpur, pp. 13 117. Pulz O. (2001). Photobioreactors: production systems for phototrophic mi croorganisms. Applied Microbiology and Biotechnology 57: 287 293. Rhee G.Y., Gotham I.J. and Chisolm S.W. (1981). Ue of cyclostat culture to study phytoplankton ecology. In: Calcott P. (ed) Continuous culture of cells. CRC Press, Florida. 159 186. Renaud S.M. and Parry D.L. (1994). Microalgae for use in tropical aquacul ture II: Effect of salinity on growth, gross chemical composition and fatty acid composition of three species of marine microalgae. Journal of Applied Phycology 6: 347 356. Renaud S.M., Luong Van Thinh, Parry D.L. (1999). The gross chemical com position and fatty acid composition of 18 species of tropical Australian mi croalgae for possible use in mariculture. Aquaculture 170: 147 159. Richmond A., Cheng Wu Z., and Zarmi Y. (2003). Efficient use of strong light for high photosynthetic productivity: interrelationships between the optical path, the optimal population density an cell growth inhibition. Biomolecular engineering 20: 229 236. Rocha J.M.S., Garcia J.E.C., and Henriques M.H.F. (2003). Growth aspects of the marine microalgae Nannochlorpsis gaditana . Biomolecular Engineering 20: 237 242. Rodolfi L., Zitteli G.C., Barsanti L., Rosati G., and Tredici M.R. (2003). Growth medium recycling in Nannochloropsis sp. mass cultivation. Biomole kular Engineering 20: 243 248.
121 Chapter 3
Schneider O., Sereti V., Eding E.H., and Verreth J.A.J. (2005). Analysis of nutrient flows in integrated aquaculture systems. Aquacultural engineering 32: 379 401. Støttrup J.G. and Norsker N.H. (1997). Production and use of copepods in marine fish larviculture. Aquaculture 155: 231 247. Støttrup J.G. and McEvoy L.A. (eds.) (2002). Live feeds in marine aquacul ture. Blackwell publishing, 336pp. Syrett P.J. (1981). Nitrogen metabolism of microalgae. In: Platt T. (ed). Physiological bases of phytoplankton ecology. Canadian Bulletin of Fisheries and Aquatic Sciences 210: 182 210. Tandler A., Mozes N., and Ucko M. (2003). The assimilation of dissolved fish waste by microalgae. Annual report 2003 of the EU project ZAFIRA (Zero discharge Aquaculture by Farming in Integrated Recirculation Systems in Asia), http://zafira.wau.nl: 42 69. Volkman J.K., Dunstan G.A., Barrett S.M., Nichols P.D., and Jeffrey S.W. (1992). Essential polyunsaturated fatty acids of microalgae used in feed stocks in aquaculture. In: Proceedings in Aquaculture Nutrition Workshop, edited by Allan G.L. and Dall W., Salamander Bay, 15 17 April 1991. NSW Fisheries, Brackish Water Fish Culture Research Station, Salamander Bay, Australia, pp. 180 186. Waller U., Bischoff A.A., Orellana J., Sander M., and Wecker B. (2003). An advanced technology for clear water aquaculture recirculation systems: Re sults from a pilot production of Sea bass and hints towards "Zero Dis charge". European Aquaculture Society Special Publications 33: 356 357. Wikfors G.H. and Ohno M. (2001). Impact of algal research in aquaculture. Journal of Phycology 37: 968 974. Von Elert E. and Stampfl P. (2000). Food quality for Eudiaptomus gracilis : the importance of particular highly unsaturated fatty acids. Freshwater Biol ogy 45: 189 200.
122 Chapter 4 Ozonation and foam fractionation used for the re moval of bacteria and particles in a marine recir culation system for microalgae cultivation Kube N. and Rosenthal H.
123 Chapter 4 Ozonation and foam fractionation used for the re moval of bacteria and particles in a marine recircu lation system for microalgae cultivation Kube N. and Rosenthal H. Abstract The use of ozone in combination with foam fractionation was tested for its effectiveness in disinfecting highly loaded waste water in a bypass of a ma rine recirculation system used for microalgae cultivation. Water was treated with ozone at different levels as measured indirectly via different redox po tentials (400mV, 500mV, 600mV). The redox levels were also applied at dif ferent retention times of the water in a foam fractionator (5min, 10min, 15min). A broad spectrum of water quality and microbial data for inflow and outflow samples was recorded: viable and total numbers of bacteria, qualita tive analysis of living and dead cells (Live/Dead® staining), size distribution of particles and amount of attached living bacteria; pH, free ozone content and dissolved nutrients of the water. The results showed clearly that ozone kills free floating bacteria under given operational conditions. However, almost all surviving bacteria were attached to mostly organic particles, where they seem to be sufficiently protected from ozone attack as revealed by the specific staining method. As anticipated by the operational criteria of the foam separation and contacting tower, much less bacteria and fine particles were detected in the outflow samples. It can be concluded the foam fractionation process contributes also substantially the mechanical removal of fines and aggregated bacteria cells.
124 Chapter 4 4.1 Introduction For several decades, interest in aquatic applications of ozone other than for drinking water sterilization (Legeron, 1984) has increased considerably (Rosenthal, 1981, Rosenthal and Wilson, 1987). Ozonation is now widely used and well established in municipal waste water treatment (Hoigne and Bader, 1980), in process water polishing (Richard and Brener, 1984), in soft drink manufacturing (Schneider and Rump, 1983), in pulp and paper waste water treatment (Rice, 1984), in swimming pool water treatment (Eichelsdo erfer and Jandik, 1985), and many other industrial applications including fish processing to extend the shelflife for fresh fish (Kötters et al. , 1997; Crappo et al. , 2004). Its application in aquaculture recirculation system has also been practiced for more than three decades (Rosenthal, 1981), mostly in experimental fish culture systems but also in a growing number of commer cial production units, particularly in hatcheries. (Rosenthal and Wilson, 1987, Buchan et al. , 2004; Sander, 1998; Summerfelt et al. , 2002; Fraser, 2004). This application gains importance (Brazil et al. , 1998; Waller et al. , 2002; Chen et al. , 2003) and has recently been increasingly studied in China (Gong et al. , 2002; Liu et al. 2003, 2004) while applying methodological principles previously developed by Rosenthal (1981). While the argument for its use was often to disinfect the water, some of the studies clearly identified the benefits of ozone use in combination with foam fractionation and nitrite oxidation. (Rosenthal, 1981; Rosenthal and Wilson, 1987; Summerfelt, 2001). The need for disinfection or even „sterilisation“ (as in drinking water applications) is not even desired in aquaculture systems as these recycling systems work with living animals, and can be considered the as a „commercial ecosystem“ in which a balanced fauna and flora of organ isms thrives (Rosenthal, 1981; Rosenthal and Black, 1993). Ozonation has, therefore, mostly been combined with foam fractionation to enhance the re moval of particles.
125 Chapter 4
It should be noted that the term „ozone“ is used here in a generic sense as ozone reacts with several seawater components to form secondary oxidants within seconds. Among ozone specialists the term TRO (total residual oxi dants) is commonly used to reflect this fact while also circumventing the dif ficulty to exactly determine the ozone concentration, particularly in seawater (Buchan et al., 2005), where some bromide radicals and hypobromous acid is formed (Tango and Gagnon, 2003). Further, the output of various mix tures of oxygen radicals in any electrical discharge ozonizer depends on sys tem configuration and operational procedures (high frequency or low fre quency ozonizer). It is also for this reason that the redox potential has been chosen as a convenient measure to characterize total radical oxidant (TRO) activity. Therefore, the common language usage of „ozone“ in this paper is meant to refer to TROs. Although ozone cannot be considered a coagulant in the classical sense, many studies have evaluated its use as a coagulant aid. Dissolved organic matter has been identified as playing a primary role in particle stabilization and hence, ozonation – through electrostatic loading and polarization of the hydrophobic hydrophilic ends foster settling characteristics of suspended solids while also assisting in forming aggregates that attach to water air interfaces, producing a stable foam which can be removed by counter current stripping. Depending on the organic and microbial load, there will certainly also be some reduction of bacterial counts, however, there have been only a few quantifying studies on the fate of bacteria in ozone contacting chambers (Wuhrmann and Meyrath, 1955), particularly with regard to virus inactiva tion (Akey and Walton, 1984; Liu et al. , 2004), however, there is very little information on the bactericidal effects of ozone application in conjunction with aquaculture recycling systems using foam fractionators. Certainly, the removal of unwanted fine particles which not only consume oxygen, but also are repeatedly in contact with respiratory tissues of the cultured species and – above all – potentially act as „carriers“ for facultative pathogens, is desir able and also reduces simultaneously the bacterial load.
126 Chapter 4
The objectives of the present study were (a) to identify the „disinfection ca pacity“ of a foam fractionation system using ozone at three different levels and three different retention (contact) times in the contact chamber and (b) to investigate whether there are differences in the survival of free floating bacteria in contrast to particle attached microbial species.
4.2 Material and Methods 4.2.1 System configuration The examined foam fractionator was a small unit operated in an extra by pass (volume 8.4 L, Fig. 1). The purpose was mainly to remove fine particles while at the same time aiming to act as a disinfection unit. The effluent of this foam fractionator was supplied to a novel photobioreactor system, in which microalgae are produced from dissolved nutrients derived from a ma rine recirculation system. The system was stocked with Sparus aurata (Kube et al., 2006a in prep., Chapter 3, Fig. 1). Considering the system configura tion the foam fractionator acted as pretreatment of fish effluents in order to remove suspended particles and bacteria entering the microalgae production unit which consisted of three acrylic columns (50L each). For detailed infor mation about the marine recirculation system and the microalgae production unit see Chapters 3 and 5, respectively. Water from the main recirculation system was continuously pumped (Eheim pump type 1260) to feed the foam fractionator. The flow rate was manually adjusted by a ball valve and flow control unit (see Table 1). Ozone containing air was supplied to the foam fractionator at a rate of 200L/min (estimate from producer table). The feed line comes from the compressed air supply, which also supports the aeration of the entire recirculation system, including water storage column. The ozone generator was a product of the Sander Company (Uetze Eltze, Germany; model S500, output characters: ozone pro duction at 500mg h 1, air flow rate 50 1000 L h 1). Air diffusers made of lime wood were used.
127 Chapter 4
Redox potential was determined in the outlet of the foam fractionator by a electrode and regulated by a control module (model KM2000, Sensortechnik Meinsberg GmbH, Ziegra Knobelsdorf, Germany) (Fig. 1, C). Foam conden sate was collected in a separate container. Waste air from foam fractionation and waste container were treated with activated carbon. Pretreated water was transferred to a storage column (Fig. 1, E) with a constant aeration for outgassing residual ozone and to stabilize the pH. Spare water from the foam fractionator was discharged via an overflow into the main water circle (Fig. 1, F). 4.2.2 Sampling methods Sampling was done at randomly chosen production days during the entire investigational period (April 2004 to March 2006). Different retention times were adjusted by a flow meter at levels according to Table 1 the day before the sampling started. This was to ensure that redox potential reached the desired level. However, we are aware that a certain flow rate in a given time does not necessarily result by definition in true retention times because of improper mixing and turbulences. However, for the purpose of this study the approximate time is considered to be adequate. Tab. 1 Overview about the adjusted flow rates for the required retention time within the foam fractionator D Retention time Flow rate/h 15 min 32 Litres C 10 min 48 Litres A B 5 min 96 Litres Fig. 1 Foam fractionator for the pretreatment of F system water. A = Inlet, B = Outflow, C = redoxpotential probe, E D=Foam collector, E = water storage column with aeration, F=overflow back to the main system
128 A Chapter 4
Samples were taken at the inlet (Fig. 3, A) of the foam fractionator and at the outflow (B) after the calculated retention time.
Chemical analysis of water samples A volume of 20ml was used to determine pH (model multi 340i, WTW, Ger many). It was frozen at 20°C for later analysis of dissolved nutrients by an Autoanalyser AA3 (Bran Lübbe, Norderstedt, Germany: Method no. G 016 91 for Nitrate N, G 029 92 for Nitrite N, G 102 93 for Ammonia N, G 103 93 for Phosphate). The content of TRO [mg L 1] was measured in outflow samples using HACH Permachem ® Reagents (DPD Free Chlorine Reagent, 10ml). Extinction was measured with a spectrophotometer (Hach, Germany) and given in free chlo rine concentrations. Values were multiplied with the factor 0.6769 to recal culate content of free total residual oxidants (from now on referred to as ozone).
Microbiological methods 1) CFU determination – viable cells Bacterial abundances were determined using the pour plate method by Koch (JAHR) with standard marine agar (TSB 3). Different dilutions were used for inflow and outflow samples (1ml to 1:1000 & 1:10.000 for inflow samples; 1:10 & 1:100 for outflow samples). Samples were incubated at room tem perature for two weeks in darkness and colony forming units (CFU) were de tected after incubation. 2) Epifluorescence Microscopy a) TBN total bacterial number Different bacterial species have different growth requirements and therefore only a small proportion of the total bacterial number will be able to grow on standard agar. To determine the total number of bacteria, the Acridine Or ange Direct Count (AODC) method was used. Acridine Orange binds to acidic cell components like DNA (orange stain, Jochem, 2001) and the stained cells
129 Chapter 4 can be counted using epiflourescence microscopy. Therefore, immediately after sampling, 20 ml of the sample were transferred into a glass vial (Perkin Elmer, PLCAP/500), fixed with 0.5ml particle free formaline and stored at room temperature in darkness for later analysis. For staining, samples were vacuum filtrated (<200 mbar) through 0.2 m black polycarbonate filters preventing background fluorescence (ø25mm) (Whatman Nucleopore). Filtration volume for Acridin Orange samples ranged from 2 ml (inflow sample) to 6 ml (outflow sample). 1 ml of acridine orange was pipetted to the sample and the dye was incubated on the filter for 5min in darkness (STOCKING). Acridine Orange was then removed by filtration ( 200 mbar); filters were air dried, embedded in oil (Cargille ®) and kept in darkness before analysis using a Zeiss Axioplan epifluorescence microscope with a filter setting KP490,KP500; Teiler 510 & LT520 (450 – 490 nm emis sion). b) Viable staining Qualitative analysis of bacterial abundances distinguishing living and dead cells was performed with LIVE/DEAD ® BacLight™ Bacterial Viability Kits for microscopy (Molecular Probes). This method includes two different nucleid acid stains differing in their spectral characteristics and in their ability to penetrate healthy bacterial cells. SYTO ®9 (3.34 mM) is green fluorescent and labels all bacteria. In contrast, propidium iodide (20 mM) penetrates only bacteria with damaged membranes, causing a reduction in the SYTO 9 stain fluorescence, when both dyes are present (Molecular Probes, 2004). So, dead cells are stained in red, whereas living cells will be stained in green. For staining, 2 ml of sample water were transferred to a 2 ml Eppendorf cup and immediately stained with both dyes (3 l). After 15min incubation at room temperature, bacterial samples were fixed with 0.07ml formaline and stored in darkness for later analysis. For analysis, 2 ml of the samples (inflow and outflow samples) were filtered in order to enrich bacteria on the filter. Filters were mounted in oil ( Bac Light
130 Chapter 4 mounting oil) and stored in the darkness. Analysis was performed by epifluo rescence microscopy analogue to the AODC method.
Calculation of TBN and viable staining Total bacterial numbers and viable bacteria were determined by counting 30 grid fields (New Porton G12; Graticules Ltd., UK) with 1000x magnification. Total bacterial numbers were calculated using the following equation: F ⋅ n = grid N (Equation 1) V where N is the total number of bacteria per ml, F is the conversion factor of 367.455 which reflects the area of the New Porton G12 grid in relation to the total filter area (inner diameter of the filter), ngrid is the average number of bacteria per grid and V the filtrated volume. Total numbers were divided by 1.000.000 to express cell concentrations in Mio cells/ml. Additionally, the size distribution of particles and the amount of bacteria liv ing attached to particles were determined by viable staining. Therefore, 10 grid countings (New Porton G12 grid, 0.44mm width, 0.74mm length each, 100x magnification) were summarized. Size classes were determined accord ing to the grid: less than <10 m, 10 50 m, 50 120 , 120 240 m, >240 m. Due to patchy distribution of the bacteria attached to particles, it was not always possible to count every single bacterium. Furthermore, only the top side of the particle was visible. Thus, bacteria attached to the back could therefore not be counted and integrated into further calculations. Neverthe less, the amount of visible bacteria attaced to particles was classified into 4 groups: 0 = none, + = 1 5, ++ = up to 20 ; +++ = >20.
131 Chapter 4 4.3 Results 4.3.1 Viable counts Viable plate counts were used as first indicator of disinfection efficiency. Ta ble 1 shows that application of TRO´s at levels of 500mV and 600mV re sulted in very low number of colony forming units (CFU) per ml, independent from the initial bacterial load of the inflow, which differs due to conditions in the recirculation system. For 600mV it was 0.600 ± 0.560 10³ CFU/ml (15min retention time in foam fractionator); 0.144 ± 0.056 10³ CFU/ml (10min retention) and 0.093 ± 0.025 10³ CFU ml 1 (5min retention), respec tively. Average numbers for 500mV were: 0.200 ± 0.149 10³ CFU ml 1 (15min retention), 0.174 ± 0.187 10³ CFU ml 1 (10min retention) and 0.039 ± 0.037 10³ CFU ml 1 (5min retention), respectively. 15min retention time at 400mV gave similar results: 0.796 ± 0.321 10³ CFU ml 1. Less retention time re sulted in a rapid increase to several thousand CFU ml 1: 6.92 ± 5.34 10³ CFU ml 1 (10 min retention time) and 23.91 ± 14.34 10³ CFU ml 1 (5min retention time).
132 Chapter 4
Tab. 2 Viable bacterial counts (pour plate method) as determined from inflow and outflow samples of foam fractionator at different redoxpotential and water retention times. Data represent means of three parallel subsamples. Total sample number for 600mV = 8, taken over two operational days. All other operational redox levels (n = 5) are taken at the same day. % dead indicates percentage decrease from inflow samples.
colony forming units (CFU) (10 3 per ml) retention 15 min 10 min 5 min time
Inflow Outflow % dead Inflow Outflow % dead Inflow Outflow % dead
600 mV 17.8 0.19 98.93 37.5 0.13 99.65 273.2 0.09 99.97 13.6 1.72 87.35 26.8 0.14 99.48 475.8 0.12 99.97 18.0 1.36 92.44 41.7 0.25 99.40 95.3 0.07 99.93 1273.3 0.45 99.96 93.5 0.11 99.88 - 14.2 0.38 97.32 27.0 0.09 99.66 - 19.5 0.12 99.38 - - 29.0 0.25 99.14 - -
25.7 0.30 98.83 - -
average 0.6 96.6 0.14 99.6 0.28 99.9
500 mV 983.3 0.49 99.95 630.7 0.51 99.92 246.7 0.03 99.99
1246.7 0.13 99.99 696.7 0.23 99.97 368.3 0.005 99.99
1333.3 0.08 99.99 480.0 0.11 99.98 201.7 0.02 99.99 1180.0 0.19 99.98 666.7 0.01 99.99 220.0 0.03 99.99 1046.7 0.11 99.99 583.3 0.01 99.99 171.7 0.11 99.99
average 0.2 99.98 0.17 99.93 0.039 99.99
400 mV 85.5 0.48 99.44 113.0 9.15 91.90 691.7 19.65 97.16
283.2 1.37 99.52 534.5 1.91 99.64 619.0 43.44 92.98
69.2 0.74 98.93 102.0 3.60 96.47 453.0 38.02 91.61 90.1 0.87 99.03 700.2 3.54 99.49 372.2 10.12 97.28 120.3 0.52 99.57 290.5 16.4 94.35 238.5 8.32 96.51
average 0.796 99.3 6.92 96.4 23.91 95.11
133 Chapter 4
4.3.2 Quantitative and qualitative analysis
Not all bacteria grow on agarplates. Staining samples with acridin orange can provide total bacteria numbers, however this method does not differenti ate between living and dead bacterial cells, so an overestimation of bacteria contamination may occur. Staining with LIVE/DEAD ® allows a qualitative analysis to distinguish between living and dead bacteria cells. The results clearly indicate that living bacteria were almost exclusively found attached to particles of different size (Fig. 3a). Unattached bacteria were al most exclusively stained red, and therefore identified as dead. Nevertheless after the ozone treatment living bacteria were still found while attached to particles (Fig. 3b), but in much smaller numbers. Both the amount of parti cles and free bacteria was much less than the inflow. Thus, most of them were removed by the foam fractionation process and were collected in the waste vessel (Fig. 3c). a b c
50 m m Fig. 3 Occurrence of living bacteria (green) attached to particles (red) (a) in the inflow and (b) in the outflow of the pretreated water. The particles are removed by the foam fractionation and can be found in the waste container (c). The examination of inflow and outflow samples showed, that removal of un attached bacteria (Fig. 4a) and particles (Fig. 4b) was very effective at every treatment level tested, even at short retention times. Similar to the results on viable counts, outflow concentration of bacteria and particles (grey bars, Fig. 4a, b) are at constantly low levels, independent of the inflow concentra tion (black bars) (ca. 1000 particles per ml and 0.5 Mio. bacteria per ml).
134 Chapter 4
4 10000 a 600 mV b 400 mV 500 mV 400 mV 500 mV 600 mV 8000 3
6000 2 4000
bacteria in Mio/ml in bacteria 1 2000 total number of particles per ml per of particles number total 0 0 15 10 5 15 10 5 15 10 5 15 10 5 15 10 5 15 10 5 retention time in foam fractionator treatment inflow outflow Fig. 4 Bacterial and particle counts in samples from the inflow and outflow of the foam fractionator. (a) bacteria (results from AO staining) and (b) total number of particles at different redox potentials (400mV, 500mV, 600mV) and retention times (15min, 10 min, 5 min). Results show low outflow con centration of both bacteria and particles (grey bars), independent from inflow concentrations (black bars). If particles play the dominant role for bacteria contamination, we have to in vestigate the question: what sizes of particles can be found and is there a bacteria preference for attaching to certain particle sizes? Size distribution of particles gave following results: At each treatment, size classes of less than 10 m and 10 50 m were the most frequent particle size. Both classes together represent at least 75%, sometimes almost 99% of the total number of the sample (Fig. 5a,b).
100 a inflow 100 b outflow
80 80
60 60
40 40
20 20 relative amount of particles amount relative of particles amount relative
0 0 15 10 5 15 10 5 15 10 5 15 10 5 15 10 5 15 10 5 400mV 500mV 600mV 400mV 500mV 600mV retention time in foam fractionator retention time in foam fractionator Fig. 5 Relative amount of particles of the size >10 m and 10 50 m in the (a) inflow samples and (b) outflow samples in relation to total number. Data represent average values of 5 replicates for each treatment at different redoxpotentials [mV]/retention time [min]. 135 Chapter 4
Regarding the abundance of particles with bacteria in relation to total parti cle numbers (Table 3, 4), the majority of particles <10 m was detected to be free of bacteria. This size class was dominated by dead bacteria forming ag glomerations (2 10 bacteria) and were classified as particles without living bacteria. In the size class 10 to 50 m, a variation between dominance of particles without bacteria and particles with attached bacteria was observed. In this size class very often eukaryotic microalgae were observed and declared as particles without bacteria (see also Tab. 4). In particle size classes 50 m to 240 m the proportion of particles with at tached bacteria increased, whereas the proportion of particles free of at tached bacteria decreased (Tab. 3). Particles >240 m were dominated by Sphaerotilus natans , a sewage bacterium forming long chains. Particles iden tified as Sphaerotilus natans were assigned to the fraction “particle without bacteria” (Tab. 3, column >240 m “0”). Sphaerotilus natans chains were mainly found in inflow samples; in outflow samples they were only detected at an redox potential of 400mV (Tab. 3, >240 m “0”). In summary, increas ing particle sizes can be correlated with increasing abundances of living bac teria attached to the particles. According to the results, particles >50 m are of minor importance for parti cle contamination because of the low abundance of these size classes. Parti cles sizes of 10 to 50 m are of major relevance for the contamination on the one hand because of their high abundance and on the other hand because of the high abundances of particle attached bacteria (Fig. 5).
136 Tab. 3 Summary of data for size distribution of particles and the amount of attached bacteria per ml sample for inflow (in) and outflow (out) samples for each treat ment (redoxpotential [mV], retention time [min], left column). Particles were sorted according to size (first row of table) and bacteria amount: 0 = no living bacteria, + = 1 5 attached bacteria, ++ = up to 20 bacteria, +++ = more than 20 bacteria. Average & SD from 5 replicates for each treatment. Total = total numbers of particles Particle size <10 m 10 50 m 50 120 m 120 240 m >240 m att. bacteria 0 + ++ +++ 0 + ++ +++ 0 + ++ +++ 0 + ++ +++ 0 + ++ +++ total 400/15 2029 153 56 181 97 28 69 14 14 2640 in ±854 ±205 ±124 ±188 ±116 ±62 ±49 ±31 ±31 1661 400/15 389 139 14 69 83 56 42 14 14 14 834 out ±349 ±203 ±31 ±85 ±76 ±58 ±38 ±31 ±31 ±31 934 400/10 208 69 69 56 153 181 97 14 28 28 14 917 in ±139 ±85 ±155 ±31 ±151 ±293 ±144 ±31 ±38 ±62 ±31 1161 400/10 125 42 111 306 56 28 14 14 14 709 out ±91 ±38 ±105 ±416 ±58 ±31 ±31 ±31 ±31 839
400/5 667 97 14 28 222 111 42 14 14 42 14 28 14 14 14 181 28 14 1556 in ±356 ±79 ±31 ±38 ±151 ±38 ±62 ±31 ±31 ±38 ±31 ±38 ±31 ±31 ±31 ±79 ±38 ±31 1166 400/5 195 28 28 167 56 42 14 42 14 14 42 639 out ±227 ±38 ±38 ±105 ±58 ±38 ±31 ±62 ±31 ±31 ±38 698 500/15 903 42 14 153 42 14 97 42 14 28 28 14 28 28 1445 in ±681 ±38 ±31 ±166 ±93 ±31 ±79 ±62 ±31 ±62 ±62 ±31 ±62 ±38 1468 500/15 514 56 28 14 14 14 639
137 out ±642 ±76 ±38 ±31 ±31 ±31 849 500/10 1792 264 459 125 28 83 167 28 28 14 42 14 3043
in ±1471 ±296 ±305 ±114 ±62 ±31 ±135 ±62 ±62 ±31 ±93 ±31 2695 500/10 452 35 226 35 35 17 799 out ±216 ±40 ±67 ±40 ±69 ±35 467 500/5 264 528 139 111 14 153 56 28 14 28 14 28 14 1389 in ±257 ±373 ±163 ±212 ±31 ±342 ±91 ±62 ±31 ±38 ±31 ±38 ±31 1700 500/5 139 361 14 14 14 14 28 14 14 611 out ±196 ±345 ±31 ±31 ±31 ±31 ±38 ±31 ±31 766 600/15 361 250 69 250 83 97 69 56 14 69 14 14 28 208 1584 in ±134 ±167 ±69 ±126 ±91 ±79 ±98 ±91 ±31 ±98 ±31 ±31 ±62 ±220 1328 600/15 153 56 83 83 14 42 56 28 14 14 542 Redoxpotential[mV]/retention time[min] out ±193 ±91 ±186 ±91 ±31 ±62 ±91 ±38 ±31 ±31 844 600/10 1158 116 69 1297 208 139 69 301 23 93 23 116 3612 in ±1300 ±145 ±69 ±2127 ±69 ±69 ±69 ±343 ±40 ±106 ±40 ±145 4454 600/10 514 56 28 14 14 14 639 out ±642 ±76 ±38 ±31 ±31 ±31 849 600/5 1945 243 556 382 139 243 139 313 69 69 69 14 35 35 69 35 4411 in ±1965 ±344 ±491 ±147 ±0 ±49 ±196 ±49 ±98 ±98 ±0 ±31 ±49 ±49 ±98 ±49 3782 600/5 1112 104 104 382 104 208 35 139 35 69 35 2327 out ±884 ±49 ±147 ±147 ±49 ±98 ±49 ±98 ±49 ±98 ±49 1719 Chapter 4
Tab. 4 Percentage of particles without (0) and with attached living bacteria (#) for each size class in inflow samples at different treatments (redoxpotential [mV]/ retention time [min], left column) in rela tion to total numbers (Table 5). Data present average values of 5 replicates at each treatment. Sum1 = sum of size class <10 m + 10 50 m; sum2 = sum of size classes 50 120 m + 120 240 m + >240 m. particle size classes 120 INFLOW <10 m 10 50 m 50 120 m >240 m 240 m treatment O # 0 # Sum1 0 # 0 # 0 # Sum2 400/15 76,8 7,9 6,8 7,4 98,3 0,5 0 0 0,5 0 0 1,1 400/10 22,7 15,2 6,1 47,0 90,9 1,5 6,1 1,5 0 0 0 9,1 400/5 42,9 8,9 14,3 10,7 76,8 0,9 5,4 0 2,7 11,6 2,7 23,2 500/15 62,5 3,8 10,6 3,8 80,8 6,7 5,8 1,9 2,9 1,9 0 19,2 500/10 58,9 8,7 15,1 5,0 87,7 2,7 6,4 0,9 0,5 1,4 0,5 12,3 500/5 19,0 0,0 38,0 19,0 76,0 11,0 6,0 1,0 3,0 2,0 1,0 24,0 600/15 22,8 20,2 15,8 15,8 74,6 3,5 5,3 0,9 2,6 13,2 0 25,4 600/10 32,1 41,0 5,8 14,1 92,9 0 3,2 0 0,6 3,2 0 7,1 600/5 44,1 26,8 3,1 15,7 89,9 0 4,7 0 2,4 0 3,1 10,2
4.3.3 Influence of ozone on efficiency of foam fractionation
High values of free ozone are necessary for sterilization (Sander, 1998). Re sults from the present study clearly showed, that bacteria are able to survive treatments of 15min at 600mV redox potential when they live attached to particles. For decontamination of a recirculation system, it is therefore of major relevance to efficiently remove particles of 10 – 50 m. Two questions are implicated based on these results: i) does the content of free ozone have an impact on the removal efficiency and ii) which redox po tential value is needed to efficiently remove particles? The solution of these problems is crucial for practical application, because technical ozone pro duction is very cost intensive. Unfortunately, the results from this study could not be affirmed statistically due to the varying conditions during the microalgae cultivation period. When single data points are plotted against the content of free ozone, evidence was obtained for an redox potential higher than 400mV appearing to be crucial for efficient particle removal. Lower retention times (< 15min.) resulted in higher viable counts (Koch method, Fig. 6a), occurrence of Spaerotilus natans (Tab. 3, >240 m “0”, out) in the outflow sample and lower particle
138 Chapter 4 removal efficiency (Fig. 6b). This effect could be alleviated by elevated reten tion times. Redox potential values > 500 mV did not yield a higher particle removal efficiency, even when retention times were varied. However, con cerning this topic, further investigations using a comparable system are needed.
50 120 a b 400mV 40 100 500mV 600mV 30 80
20 60
10 40 CFU in 10-3 per ml ml per 10-3 in CFU relative removement relative 0 20
0 0,0 0,2 0,4 0,6 0,8 1,0 1,2 1,4 1,6 0,0 0,2 0,4 0,6 0,8 1,0 1,2 1,4 1,6
free ozone content in mg/l content of free ozone [mg/L] Fig. 6 Free ozone content for each treatment (5min, 10min and 15min. retention time) at total residual oxidant levels (TRO) in mgL 1. (a) colony forming units, (b) relative amount of removed particles from inflow (data are shown for values with at least 2000 particles per ml inflow).
4.4 Discussion Because of its known disinfection capacity, ozone is believed to be most ef fective to reduce the bacterial load of the water in aquatic applications (Sander, 1998). While this is true for drinking and process water sterilization where high dosages are employed, the use of ozone in aquaculture has never been exclusively targeted on this goal because of the risk to damage the cul tured species at the same time (Brazil et al. , 1998). Therefore the application mostly aims at a multiple function with: (a) primarily enhanced removal of fine particles through aggregation (e.g. electro static loading) and counter current foam stripping, (b) simultaneous oxidation of organic compounds with multiple double bonds that are not easily degradable in biofilters (e.g. conversion into bio degradable breakdown products to be returned to the biofilter) and finally (c) to reduce to some extent the overall microbial load (Rosenthal, 1981; Liltved, 2000; Summerfelt et al. , 2002).
139 Chapter 4
The interest to understand the qualitative and quantitative effects of by pass ozonation on overall reduction in pathogen counts has recently been revived in relation to recirculating aquaculture systems to combate fish diseases such as Saprolegniasis in freshwater systems (Forneris et al. , 2003) and ma rine halibut rearing facilities (Fraser, 2004) as well as shrimp hatcheries (Meunpol et al. , 2003). Bullock et al. (2002) clearly demonstrated that num ber of pathogen counts can be reduced in freshwater recirculation systems to reduce the risk of disease outbreaks such as bacterial gill diseases (BGD). While ozonation did reduce BKD mortality in fish, it failed in nearly all cases to produce even a one log 10 reduction in numbers of heterotrophic bacteria in the system water or on gill tissue. Failure of the ozone to lower numbers of heterotrophic bacteria significantly or to prevent the causative BGD bacte rium from occurring on gills was attributed to the short exposure time to ozone residual (35 s in contact chamber) and rapid loss of oxidation effi ciency caused by levels of total suspended solids and nitrite. Our study clearly shows that retention time and thereby contact time, has certainly a major role in effectiveness to damage heterotrophic bacteria, due to increas ing levels of TROs. The amount of total residual oxidants plays a major role for survival of bacterial fish pathogens. Sugita et al. , 1992, showed that bac terial counts of Enterococcus seriolicida , Vibrio anguillarum and Pasteurella piscicida decreased by more than 0.040 to 0.060 mg of total residual oxi dants per litre, whereas no decrease in viable counts was observed at less than 0.018 to 0.028mg of TROs per litre. Similar results have been found in this study. Hence, the effectiveness of ozone treatment depends on TRO con centration, length of ozone exposure (contact time), pathogen loads and lev els of organic matter. However, it seems not advisable to apply ozone directly in the main stream of a fish culture recirculation system, because high levels of ozone can have a negative effect on fish health. Ozone appears to cause severe damage to all external tissues. This can result in disruption of respiration, osmoregulation and possibly excretion, resulting either in death or in the development of various sublethal secondary pathological effects (Paller and Heidinger, 1980).
140 Chapter 4
But it is possible to introduce ozone in a by pass stream which receives only a portion of the total flow. When this by pass flow is returned to the main stream, the residual total radical oxidant concentration is drastically re duced, thereby greatly reducing the risk of toxic concentrations entering the fish culture tank. Further, residual ozone in the effluent of the counter cur rent foam stripping unit will immediately react with organic particles and other reactants in the main stream, thereby quickly reducing the remaining TROs to safe levels. Otherwise this study showed that use of foam fractionation supports the dis infecting process, as most of the living bacteria were determined to be at tached to particles. Particles of the size less than 10 m and between 10 to 50 m turned out to be the major size class because of their major abun dance in total numbers of suspended solids. This has been concluded also in other studies (Orellana et al. , 2005). One third of the solid load of a recircu lation system can be found as suspended solids (Waller et al. , 2003a). High numbers of small particles provide a large surface area resulting in sufficient substrate for bacteria. This fraction needs to be removed, as to our knowl edge this study documents for the first time by directly distinguishing be tween life and dead bacterial cells that bacteria can even survive treatments at high TROs at retention time of 15 minutes, when being attached to parti cles. Especially the surfaces of particles can provide microhabitats for bacte ria (Curds, 1982; Acinas, 1999). The particles can serve as food source (i.e. in the marine realm, particle attached bacteria are thought to play an impor tant role in carbon cycling (Cammen and Walker, 1982; Irriberry et al. , 1990) but also provide protection from bacterivores (Curds, 1982). Even suboxic/anoxic conditions may develop and hence allow anaerobic bacteria to survive at these surfaces. In marine snow, several bacterial species are known to produce a capsular envelope in order to ascertain suitable condi tions and to protect themselves against bacterivores (Heissenberger et al. , 1996). There is one advantage of ozone application over UV light use in recircula tion systems. While UV disinfection targets on DNA destruction, ozone (total
141 Chapter 4 radical oxidants) reacts with the surfaces of microbial membranes. While UV light may only cause partial damage to bacterial cells and may enhance mutation (e.g. UV resistant strains), ozone application is an all or none func tion. Either membrane double bonds are broken up by ozonlysis, or the membrane stays intact, partly being protected by surface contacts with par ticles, thereby permitting full survival of the cell. For application purposes in aquaculture, a two step solid separation (sedi mentation, foam fractionation) is needed to remove all size classes of solids and has been recommended before to achieve sufficient hygienic conditions in farming systems (Waller et al. , 2003a,b). Ozonation improves removal of total suspended solids in foam fractionators due to a decrease in particle stability (Rueter and Johnson, 1995). 4.5 Conclusion The present study indicates that effective particle aggregation and removal can be achieved with relatively low ozone levels. At high suspended solid loads it seems not advisable to apply very high dosages as this will not achieve noticeable simultaneous disinfection as most bacteria will survive in attachment to individual particles and particle aggregates. However, even without killing bacteria, the overall bacterial counts will be reduced mainly as an effect of removal of large amounts of fine particles to which these bac teria adhere. The „disinfection“ is therefore indirect. 4.6 Acknowledgements We thank Regine Koppe and Hans Georg Hoppe (IFM GEOMAR) for support ing this study.
142 Chapter 4 4.7 References Acinas SG, Anton J, and Rodriguez Valera F (1999). Diversity of Free Living and Attached Bacteria in Offshore Western Mediterranean Waters as De picted by Analysis of Genes encoding 16S rRNA. Applied and Environmental Microbiology 65 (2): 514 522. Akey D.H. and Walton T.E. (1985). Liquid phase study of ozone inactivation of Venezuelan equine encephalomyclitis virus. Appl. Environ. Microbiol. 50 (4): 882 886. Brazil B.L., Libey G.S., Coale C.W., and Boston, H.L. (1998). Economic analysis of hybrid striped bass (Morone chrysops x Morone saxatilis) produc tion in ozonated and non ozonated pilot scale recirculating aquaculture sys tems. Aquaculture '98, (World Aquaculture Society, Book of Abstracts, An nual Conference, Las Vegas, NV (USA), 15 19 Feb 1998, 646 pp.
Brazil B.L., Summerfelt S.T., and Libey G.S. (2002). Application of Ozone to Recirculating Aquaculture Systems. 1. International Conference on Recircu lating Aquaculture, Roanoke, VA (USA), 19 21 Jul 1996, Monograph document code VSGCP C 00 001, Virginia, USA. Buchan K.A.H., Martin Robichaud D.J, Benfey T.J., MacKinnon A. M., and Boston L. (2004). The efficacy of ozonated seawater for surface disinfection of haddock (Melanogrammus aeglefinus) eggs against piscine nodavirus. Aqua culture Association of Canada Special Publication 8: 30 33. Buchan K.A.H., Martin Robichaud D.J., and Benfey T.J. (2005). Measure ment of dissolved ozone in sea water: A comparison of methods. Aquacul tural Engineering (Aquacult. Eng.) 33(3): 225 231. Bullock G.L., Summerfelt S.T., Noble A.C., Weber A.L., Durant M.D., and Hankins J.A. (2002). Effects of Ozone on Outbreaks of Bacterial Gill Disease and Numbers of Heterotrophic Bacteria in a Trout Culture Recycle System 1. International Conference on Recirculating Aquaculture, Roanoke, VA (USA), 19 21 Jul 1996. Monograph document codeVSGCP C 00 001;Virginia, USA. Cammen L.M. and Walker J.A. (1982). Distribution and activity of attached and free living suspended bacteria in the Bay of Fundy. Can. J. Fish. Aquat. Sci. 39: 1655 1663. Chen C., Wooster G.A., Getchell R.G., Bowser P.R., and Timmons M.B. (2003). Blood chemistry of healthy, nephrocalcinosis affected and ozone treated tilapia in a recirculation system, with application of discriminant analysis. Aquaculture 218 (1 4): 89 102. Crappo C., Himelbloom B., Vitt S., and Pedersen L. (2004). Ozone Efficacy as a Bactericide in Seafood Processing. Journal of Aquatic Food Product Tech nology (J. Aquat. Food Prod. Technol.)13 (1): 111 123.
143 Chapter 4
Curds CR (1982). The ecology and role of protozoa in aerobic sewage treat ment processing. Annu. Rev. Microbiol. 36: 27 46. Eichelsdörfer D. and Jandik J. (1985). Long contact time ozonation for swimming pool water treatment. Ozone: Sci. Eng 7 (2): 93 106. Forneris G., Bellardi S., Palmegiano G.B., Saroglia M., Sicuro B., Gasco L., and Zoccarato, I. (2003). The use of ozone in trout hatchery to reduce sapro legniasis incidence. Aquaculture 221 (1 4): 157 166. Fraser K.B. (2004). Using ozone in a recirculating aquaculture system for Atlantic halibut (Hippoglossus hippoglossus): Water quality, toxicity, and economic considerations. Master thesis , Dalhousie University (Canada); Masters Abstracts International (Masters Abst. Int.) . 43 (2) pp. 511 (ISBN: 0612941132). Gong X., Liu Q., Wang Q., Li J. (2002). Study on the application of ozone in incubation of Artemia salina cysts. Marine sciences/Haiyang Kexue (Mar. Sci./Haiyang Kexue). 26 (6): 68 71. Heissenberger A., Lepperd G.G., and Herndl G.J. (1996). Ultrastructure of marine snow. II. Microbiological considerations. MEPS 135: 299 308. Hoigne J. and Bader H. (1975). Ozonation of water. Role of hydroxyl radicals as oxidizing intermediates. Science 190 (4216): 782 784. Irriberry J., Unanue M., Ayo B., Barcina I., and L. Egea (1990). Bacterial production and growth rate estimation from [ 3H]thymidine incorporation for attached and free living bacteria in aquatic systems. Appl. Environ. Micro biol. 56: 483 487. Kötters J., Prahst A., Skura B., Rosenthal H., Black E.A., and Rodigues Lopez J. (1997). Observations and experiments on extending shelf life on 'rockfish' (Sebastes spp.). J. Appl. Ichthyol. 13: 1 8. Legeron, J.P. (1984). Ozone disinfection of drinking water. In: Rice R.G., Net zer A. (eds). Handbook of Ozone technology and Applications. Vol. 2 Ozone for Drinking water treatment. Butterworth:Boston, MA,USA, pp. 1 20. Liltved H. (2000). Disinfection of water in aquaculture: Factors influencing the physical and chemical inactivation of microorganisms. 40 pp. PhD. The sis, Norwegian College of Fishery Science , University of Tromsoe (Norway).
Liu Q., Li J., and Gong X. (2003). Study on the toxicity of ozone to different development stage of Penaeus chinensis . Shandong fisheries/Qilu Yuye (Shandong Fish./Qilu Yuye). 20 (9): 18 19.
144 Chapter 4
Liu Q., Li J., Gong, X., and Wang Q. (2004). Efficiency of ozone on three kinds of bacteria of Vibrio in different media. Marine fisheries re search/Haiyang Shuichan Yanjiu (Mar. Fish. Res./Haiyang Shuichan Yan jiu) 25 (2): 77 82. Meunpol O., Lopinyosiri K., and Menasveta P. (2003). The effects of ozone and probiotics on the survival of black tiger shrimp ( Penaeus monodon ). Aquaculture 220 (1 4): 437 448. Orellana J., Wecker B., Sander M., and Waller U. (2005). Particulate matter in a modern marine recirculation system: what, where and how much. European Aquaculture Society Special Publication 35: 354 355. Paller M.H. and Heidinger R.C. (1980). Mechanisms of delayed ozone toxicity to Bluegill Lepomis machrochirus Rafinesque. Environmental Pollution (Se ries A) 22: 229 239. Pirt S.J. (1975). Principles of microbe and cell cultivations. Blackwell Scien tific publications. 284pp. Richard Y. and Brener L. (1984). Several articles In: Rice R.G. and Netzer A. (eds). Handbook of ozone technology and applications. Vol.2 Ozone for drink ing water treatment. Butterworth Publ.: Boston MA, USA. Rice R.G. and Netzer A. (eds). Handbook of ozone technology and applica tions. Vol.2 Ozone for drinking water treatment. Butterworth Publ.: Boston MA, USA, 378 pp. Rosenthal H. (1981). Ozonation and sterilization. In: Tiews K. (ed). Aquacul ture in heated effluents and recirculation systems. Heenemann Verlag: Ber lin, Germany. Schriften der Bundesforschungsanstalt für Fischerei, vol.16/17. Rosenthal H. and Wilson J.S. (1987). An updated bibliography (1845 1986) on ozone, its biological effects and technical applications. Canadian Techni cal Report of Fisheries and Aquatic Sciences No. 1542. 249pp. Rosenthal H. and Black E.A. (1993). Recirculation systems in aquaculture. Pp. 284 294 In: Wang J.K. (ed.). Techniques for modern aquaculture. Pro ceedings of an Aquacultural Engineering Conference, Spokane, Washington, USA, June 1993. (American Society of Agricultural Engineers. 604 pp. Rueter J. And Johnson R. (1995). The use of ozone to improve solids removal during disinfection. Aquacultural Engineering 14: 123 141. Sander M. (1998). Aquarientechnik in Süß und Seewasser. Ulmer Verlag. 256pp. Schneider W. and Rump H.H. (1983). Use of ozone in the technology of bot tled water. Ozone:Sci.Engng 5 (2): 95 101.
145 Chapter 4
Summerfelt S.T., Hankins J.A., Weber A.L., and Durant M.D. (2002). Effects of Ozone on Microscreen Filtration and Water Quality in a Recirculating Rainbow Trout Culture System. 1. International Conference on Recirculating Aquaculture, Roanoke, VA (USA), 19 21 Jul 1996. Monograph document codeVSGCP C 00 001;Virginia, USA. Summerfelt S.T., Sharrer M.J., Hollis J., and Gleason L.E. (2004). Dissolved ozone destruction using ultraviolet irradiation in a recirculating salmonid culture system. Aquacultural Engineering 32 (1): 209 223. Sugita H., Asai T., Hayashi K., Mitsuya T., Amanuma K., Maruyama C., and Deguchi Y. (1992). Application of ozone disinfection to remove Enterococcus seriolicida , Pasteurella piscicida , and Vibrio anguillarum from seawater. Appl. Environ. Microbiol. 58 (12): 4072–4075.
Tango M.S., Gagnon G.A. (2003). Impact of ozonation on water quality in marine recirculation systems. Aquacultural Engineering 29 (3 4): 125 137. Waller U., Schiller A., Orellana J.,and Sander M. (2002). The growth of young seabass (Dicentrarchus labrax) in a new type of recirculation system. ICES CM 2002/S:06. Waller U., Attramadal K., Koppe R., Orellana J., Sander M., Schmaljohann R. (2003a). The control of hygienic conditions in seawater recirculation sys tems: the use of foam fractionation and ozone. European Aquaculture Soci ety Special Publication 33: 354 355. Waller U., Bischoff A.A., Orellana J., Sander M., and Waller U. (2003b). An Advanced technology for clear water aquaculture recirculation systems: re sults from a pilot production of sea bass and hints towards „zero discharge“. European Aquaculture Society Special Publication 33: 356 357. Wuhrmann K. and Meyrath J. (1955). The bactericidal action of aqueous so lutions of ozone. Schweiz. J. Path. 18: 1060 1069.
146
Chapter 5 MARE – Marine Artificial Recirculated Ecosystem II: Influence of the nitrogen cycle in a marine recirculation system with low water dis charge by cultivating detritivorous organisms and phototrophic microalgae Kube N., Bischoff A.A., Blümel M., Wecker B. and Waller U. (2006)
147 Chapter 5 MARE – Marine Artificial Recirculated Ecosystem II: Influence on the nitrogen cycle in a marine recirculation system with low wa ter discharge by cultivating detritivorous or ganisms and phototrophic microalgae Kube N., Bischoff A.A., Blümel M., Wecker B. and Waller U.
Abstract Water as well as nutrient recycling gained more and more attention in recir culating aquaculture systems during the last few years. The newly developed recirculating system MARE (Marine Artificial Recirculating Ecosystem) showed great potential towards these requirements. The experiment MARE II focused on the integration of a detritus removing compartment using Nereis diversicolor (Polychaeta) and the development and integration of a microalgae bioreactor into the system. Solids derived from the particulate fish waste were driving growth and repro duction of Nereis diversicolor . One complete life cycle of the polychaete was observed under artificial conditions within a period of 110 days. Continuous culture of the microalgae Nannochloropsis spec. ( max = 0.025 h 1) was devel oped and integrated as secondary bypass system for the production of valu able microalgae nutrition. The consumption of organics and the bioturbation effect of the worms kept the organic load of the sediment constant over a period of about 90 days. Due to the monotelic reproduction of the worm species larger individuals of Nereis diversicolor disappeared after spawning. The decreased food consump tion led to an accumulation of organic material in the first section of the de tritivorous reactor. This increased organic content created anoxic conditions leading to enhanced denitrification. Decreasing nitrate concentrations and increasing pH values in the system were the result of this development. Even though bacterial activity at all nitrifying components (biofilter, sedi ment surface, system walls, etc.) could be measured, total ammonia nitrogen (TAN) concentrations were increasing. These elevated NH 4 concentrations are assumed to be caused by another source than fish metabolism. Potential ex planations for these elevated concentrations may be the reduced nitrification rate due to elevated denitrification rates or the assimilatory nitrate reduction processes both occurring when suboxic/anoxic conditions are present. Keywords: marine recirculation system, integration, Nereis diversicolor , Sparus aurata , Nannochloropsis spec., artificial ecosystem
148 Chapter 5 5.1 Introduction The integration of secondary biological treatment components in marine re circulation systems is gaining more importance in order to reduce the envi ronmental impact of aquaculture systems (e.g. eutrophication, Chopin et al., 2001). Cultivation of additional organisms within these systems may en hance their profitability (Chopin et al., 2003; von Harlem, 2006). However, the integration of additional organisms requires considerable financial in vestment during the setup. Besides new developments in conventional production units (Neori et al., 2000), the first concept of a marine recirculation system with water and nu trient recycling was realised by integration of macroalgae filtration and ma rine ragworms (MARE, see Chapter 2). It could be shown that detritivorous organisms like Nereis diversicolor are suitable secondary organisms for utili zation of suspended solid waste from fishtanks (Bischoff, 2003). Macroalgael filters are the typical secondary treatment components for mariculture (Neori et al. , 2000; Chopin et al. , 2003). Nevertheless, integra tion of additional steps needs to be considered to reduce impacts of maricul tures on the environment. Potential candidates are bivalves, shrimps, sea cucumbers and different microalgae, but only few attempts have been con ducted to investigate the effectiveness of these organisms as additional steps in mariculture systems (Sphigel et al. , 1993; Neori et al., 2000). The main task in marine recirculation aquaculture systems (RAS) is the re moval of dissolved nutrients from the recirculated water. To date, there are almost no alternatives to bacterial or macroalgae filters for removal of dis solved nutrients. The use of microalgae filters was not tested intensively ex cept for some studies mainly focused on diatoms (Hussenot, 2003; Tandler, 2003). However, suspended microalgae like Nannochloropsis , Rhodomonas , Tetraselmis etc. have a high value for aquaculture feed, e.g. for shrimps, her bivorous organisms (Rotatoria, Copepods) or larviculture (Sargent et al., 1997; Brown et al. , 1997; Bessonart et al. , 1999). The present study was per
149 Chapter 5 formed in order to evaluate a novel continuous photobioreactor system with an automatic water pretreatment and harvesting process for the cultivation of microalgae as part of a recirculating aquaculture system. 5.2 Material and Methods 5.2.1 Modifications of the recirculation system The system configuration of the investigated recirculation system is pre sented in Wecker et al. , 2006 (Chapter 2). Here, the first version of a low wa ter discharge multi loop recirculation system MARE (Marine Artificial Recir culation System, 4.5 m³) is described. A modification of this system was also used in the present study (Fig. 1). Fish biomass was increased compared to MARE I; each fish tank (1) (700L each) was stocked with 35 Gilthead seabream (Sparus aurata ) and 65 additional animals were cultivated in the former macroalgae tank (3, 1500L). Average size of fish in all tanks was 354.9 ± 48.8g, resulting in an initial stocking density of 9.8 kg per m³ system volume at the start of the experiment. The removal of dissolved nutrients was achieved by means of a photobioreac tor system for the continuous cultivation of microalgae (for details see Chap ter 3). This system consisted of a disinfection unit, a production unit and a harvesting unit (see also Chapter 3). In the disinfection unit water was pre treated in a foam fractionator at high redox potential values (500 600mV, 0.8 1.0mg L 1 total residual oxidants, TRO). Water was afterwards trans ferred to a degassing tower with compressed air aeration to remove residual ozone (B). Further details of the “disinfection” and harvesting unit are out lined in Chapter 4 and Chapter 3, respectively. The cultivation unit for microalgae consisted of three acrylic columns (20cm diameter, 1.50 m height) with Y shape airdiffusers and an attached light tubes. Nannochloropsis sp. was cultivated and continuously harvested by a second foam fractionator. The foam condensate carrying the microalgae was collected in a separate tank. Water treated this way then flowed back into the main water circulation system at the outflow of tank 3.
150 Chapter 5
In the modified system (MARE II) a trickling biofilter (7) was installed to en hance nitrification in order to maintain minimal concentrations of ammonia and nitrite. The bioactive surface of the biofilter was approx. 35m². Several units were not modified for application in MARE II (Fig. 1): the detri tivorous culture tank (2), two foam fractionators (5) and pump (4) were used analogous to MARE I. Surface area of the detritivorous culture tank was 2.08 m². Two PVC walls were integrated to enhance particle subsidence in the tank (Waller, 2000). The tank was filled with a 10cm thick layer of sand (grain size ≤ 2mm) providing the preferred sediment for the selected worm species. Foam fractionators (Erwin Sander Elektroapparate GmbH, Outside Skimmer III) were rinsed automatically with a secondary fresh water loop, which is not illustrated in Fig. 1. The water circulation was driven by a cen trifugal pump (Argonaut G8) supplying fishtanks (700 800L h 1), foam frac tionators (1000L h 1) and biofilters (500 1000 L h 1), respectively. The detri tivorous bioreactor was restocked with a second generation of Nereis diversi- color, at approx. 850 individuals per m².
2 1 1
7 4
6 3 5
Fig. 1 Flow chart of the MARE system II (Marine Artificial Recirculating Ecosystem). Fishtanks = (1, 3), detritivorous culture tank = (2), pump = (4), foam fractionators = (5), microalgae cultivation system =(6) and trickling biofilter = (7). Double triangle = tap. Dimensions of the single modules are given in Chapter 2 and 3.
151 Chapter 5
5.2.2 Measurements The duration of the experiment was 5 months, lasting from 5 th of September 2005 to 15 th of February 2006. Biomass determination According to the average fish weight, fish were fed with pellets of 4.5 and 6 mm diameter, respectively (Biomar, Aqualife 17). The nutrition type was changed from trout feed to salmon feed after 3 months. Daily feeding rate (% fish body weight) was adjusted according to biomass determination of the fish. Feeding rate was changed from 1.0% to 0.8%, when the fish weight reached 500g. Worm biomass was determined using four subsamples. Therefore, the detri tivorous tank was divided into four areas and a sediment core of 9.5 cm di ameter (800 cm 3) was sampled from each area. Sediment was removed using a sieve of 1 mm meh size collecting the worms. Measured average weight and stocking density of collected worms were used to calculate biomass parame ters. Optical density measurements were used to determine algal biomass in the bioreactors. Therefore, at the start of the experiment, a calibration was es tablished in order to correlate algal biomass with measured optical density. Cells were counted using a light microscope and a haemocytometer (Fuchs Rosenthal). Optical density of the sample was determined at 665 nm using a HACH SR2010 photometer. During the experiment, algae were harvested by foam fractionation. The amount of algae in the harvesting unit was recorded using optical density measurements. 12 subsamples of the harvest were taken as replicates and centrifuged at 5000rpm for 10 minutes. Analysis of these samples is de scribed in the subsection Solid components (see below).
152 Chapter 5
Water chemistry Water samples for chemical analysis of dissolved nutrients were taken daily at four definite points of the MARE system: outlet fishtanks, outlet Nereis tank, outlet fishtank 3 and outlet biofilter.
Water samples were stored at –20°C for later analysis. Concentration of PO 4
P, NO 3 N, NO 2 N and TAN were analysed using an Autoanalyzer 3 (Bran Lübbe, Norderstedt, Germany). Flow rates through fish tanks and foam fractionators were recorded and ad justed to 800L h 1 and 1000L h 1, respectively. Flow rates of the microalgae reactors were recorded as well. Online measurements of ozone reduction potential (ORP), pH and dissolved oxygen were recorded with an electronic control module (KM 2000, Meins berg) and a portable measuring device (WTW multi 350). Water level of the system was controlled daily and if necessary, adjusted with filtered brackish seawater. Salinity of the system water was 24.8±1.0 psu; artificial sea salt was used to increase salinity if necessary. The foam collec tors were cleaned daily and the freshwater for rinsing the foam collectors was changed. The daily loss of water via the foam fractionators was recorded by the increased water volume in the secondary freshwater loop tank.
Solid components Samples of rinsing water of the foam fractionator were taken weekly in order to determine the amount of suspended solids removed by the system. 12 tubes of 10ml sample volume were centrifuged and the supernatant was stored at 20°C for later water analysis of dissolved inorganic nutrient load. Dry matter and C/N ratio of the centrifuged pellets were determined as de scribed below. Organic matter of the sediment was analysed weekly by the incineration technique using 5 subsamples.
153 Chapter 5
For analysis of solid waste, harvested microalgae, tissues and sediment pa rameters the incineration technique was used as well: dry matter content was determined by dehydration of the sample in a drying furnace at 60°C overnight. Organic content was measured by incineration of organic matter in a muffle furnace. C/N ratio was determined by gas chromatography (GC) in an element analyser (EURO EA elemental analyser, Milano, Italy). Energy content was measured by complete sample combustion using an IKA calo rimeter C4000. Weighing was performed using a Sartorius A 210 P (max. 200g) and a Sartorius U 4600 P (max. 4000g). 5.3 Results The MARE experimental phase I showed a good performance of the system for the cultivation of Sparus aurata by integration of biological secondary steps ( Nereis diversicolor , Solieria chordalis ). A second experimental phase (MARE II) combining fish, ragworms and microalgae was started on 5 th of September 2005 and was finished on 15 th of February 2006. This period represents experimental days 480 to 643 of the whole MARE system (Chap ter 2), but will here be referred to as “experimental days 1 163”. 5.3.1 Module fish Due to the increasing individual size of fish it was expected that fish growth switched from exponential to linear growth according to the Bertalanffy growth equation (Bone & Marshall, 1985). This hypothesis could be con firmed by experimental data; relevant data are presented in Fig. 3. Therefore, growth can be described by a linear regression: = + ⋅ f y0 a x (Equ. 1)
154 Chapter 5
700
600
500
400 average fishweight [g]fishweight average fishweight [g]
300
0 50 100 150 experimental days Fig. 3 Growth performance of Sparus aurata during the second MARE phase. Linear regression gave a= 1,60; y 0= 418, r²=0.68. 5.3.2 Module detritivorous tank Within 60 days after stocking, worms achieved a maximum weight of 0.62 ± 0.27g. Total worm biomass within the reactor was approx. 1.7 kg (Tab. 1, Fig. 5) with a standing stock density of approx. 1300 1400 individu als/m². Variations in stock density values are assumed to be caused by in accuracies of the sampling method. Tab. 1 Growth performance and biomass data of Nereis diversicolor during MARE II. aver. worm weight stocking den- total Experimental date ± SD sity ± SD biomass day [g] [ind./m²] [g] 10.10.2005 43 0.160 ± 0.259 895 ± 345 299 01.11.2005 65 0.243 ± 0.236 1011 ± 423 512 24.11.2005 88 0.369 ± 0.146 1371 ± 742 1057 15.12.2005 109 0.617 ± 0.273 1299 ± 120 1674 07.01.2006 132 0.584 ± 0.295 1112 ± 288 1356 26.01.2006 151 0.311 ± 0.329 953 ± 658 618 14.02.2006 170 0.240 ± 0.279 1631 ± 1696 818 The worm nutrition was solely based on the particulate matter from the fish tanks of the recirculating system. Tab. 2 gives an estimate of the food energy supplied to the worms at different time intervals during the course of the ex 155 Chapter 5 periment. The energy content ranged from 500 to 630 kJ d 1. Assuming that 1400 worm individuals inhabit one square meter of sediment in the detritivo rous reactor, approximately 162 J d 1 were available for an individual worm. Tab. 2 Calculated total energy per day available for Nereis diversicolor in the bioreactor according to the fish feeding rate per day. Calculations were performed according to MARE I (Chapter 2). Time interval Fish feed daily solid waste Daily total Daily total per day per day ± SD organic load ± energy ± SD [g] [g] SD [kJ] [g] 23.09. – 13.10.05 262.4 38.8 ± 9.2 25.0 ± 6.0 567.4 ± 135.1 13.10. – 08.11.05 292.0 43.1 ± 10.3 27.8 ± 6.6 631.4 ± 156.4 08.11. – 06.12.05 274.6 40.6 ± 9.7 26.2 ± 6.2 593.8 ± 141.6 06.12. – 09.01.06 232.6 34.6 ± 8.2 22.2 ± 5.3 503.0 ± 119.9 09.01. – 27.01.06 265.0 29.1 ± 9.3 25.3 ± 5.3 573.0 ± 136.4 27.01. – 15.02.06 285.2 42.1 ± 10.0 27.2 ± 6.5 616.6 ± 147.0
a 1,5 <0.15 g <0.30 g <0.45 g <0.60 g <0.75 g <0.90 g >0.90 g
1,0 0 1 rel. abundance
b
0,5 average wormweight [g] wormweight average
Verteilung 623 0,0
40 60 80 100 120 140 160 180 experimental day Fig. 5 Growth performance of Nereis diversicolor during MARE II (a) relative size distribution (weight classes) of every biomass determination. (b) average worm weight ± SD of each biomass determination (according to Tab. 1). The data indicate the generation cycle with reproduction around exp. day 109 and occurrence of new generation at day 132. The growth performance of the worms is summarized in Fig. 5. The average individual weight at the beginning of the experiment was 0.160 ± 0.259 g.
156 Chapter 5
Average individual worm weight increased until day 109 and slightly de creased at day 132 as shown in Fig 5 b). Maximum worm weight was be tween 0.617 ± 0.273 g and 0.584 ± 0.295 g. The average individual weight decreased substantially at day 151 and 170 to 0.311 ± 0.329 g and 0.240 ± 0.279 g, respectively. Fig. 5 a) shows the size distribution for the collected worms. A dramatical population decrease could be observed between day 120 and day 150. Sexual maturity of the worms is indicated by a colour change from red to a greenish colour (Hartmann Schröder, 1996), a colour change of the worms was observed for the first time on experimental day 60. Currently, no data are published concerning spawning weight of Nereis diversicolor. Data ob tained during this study show, that worm body weight at time of maturation was approximately 0.8 g. Due to the monotelic reproduction mode (death af ter spawning), large individuals disappeared after spawning from the detri tivorous culture tank. After the reproduction event, a biomass decrease could be observed within the system. Reproduction was observed for a period of approx. 20 days of the experiment. Individuals of the new generation were captured at day 100 for the first time, the start of the new “growing” genera tion can be assumed for day 120. Reproduction could not be controlled but had an enormous influence on the system stability. Due to growth of fish larger amounts of suspended particles were transferred into the detritivorous tank. This increased particle load ap parently was not completely used by the decreased worm biomass after day 132 leading to an increase of organic matter on the sediment in the first sec tion of the wormtank at the water inlet. An increase of organic load was found at the first sampling point in the reactor. Further sampling points within the Nereis tank did not show such an increase (Fig. 6). The increased amount of organic matter in this reactor probably led to a cas cade of effects within a short time interval from day 100 on: oxygen condi tions in the first section of the detritivorous reactor changed from oxic to
157 Chapter 5 suboxic/anoxic. Under these specific conditions, the process of denitrifica tion is favoured, indicated by a rapid decrease of nitrate concentrations (Fig. 6b). Denitrification is a process releasing hydroxyl ions and consequently, pH increases. This could be experimentally confirmed by the pH measure ments. Thus, to maintain system stability, a constant acid addition to regu late pH was required (Fig. 6d). Concurrently, TAN concentrations rapidly rose up to 10 mg L 1 (Fig. 6c). Flow rate through the biofilter was increased to 1000 L h 1 for the remaining experimental period and 1000 litres of water were exchanged in order to stabilize TAN concentrations. During MARE II, no loss of fish due to ammonia toxication could be observed.
10 100 a b 8 80
6 60 -N [mg/l] -N 4 3 40 NO organic content [%] content organic 2 20
0 0 40 60 80 100 120 140 160 180 0 20 40 60 80 100 120 140 160 180 experimental days experimental days
9,0 10 d c 8,5 8 8,0 6 7,5
4 values pH 7,0
TAN[mg/L] 2 6,5 0 6,0 0 20 40 60 80 100 120 140 160 180 0 20 40 60 80 100 120 140 160 180 experimental days experimental days Fig. 6 (a) organic content of the sediment within the detritivorous culture tank: Amount of organic mat ter at sampling point 1 (closed circles) increased significantly during the experimental period (ANOVA, p<0.001). Open circles are all other sampling point in the detritivorous reactor. (b) NO 3 N concentrations in the system water during MARE II accumulation of nitrate occurred until day 100. Afterwards concen trations of NO 3 N decreased to less than 10mg/L 1 due to denitrification; (c) TAN concentrations of the system water. d) pH changes.
158 Chapter 5
5.3.3 Module microalgae bioreactors Photobioreactors for the continuous cultivation of Nannochloropsis sp. were included into the MARE system. It was possible to cultivate Nannochloropsis in a continuous culture based on dissolved nutrients derived from the recir culation system. The daily harvested yield of algae biomass was 2.49 ± 1.34g DW d 1 in average. Harvest was very efficient: daily harvest volume was 0.184 ± 0.053 L d 1 at high cell densities of 227.9 x 10 8 ± 6.7 x 10 8 cells ml 1. 99.68 ± 0.38% of the water coming from the photobioreactors were trans ferred back to the main recirculation system. Nutritional value of the algae was high: organic proportion reached more than 80% of the algae dry weight, measured energy value was 21.09 ± 4.6 KJ g 1 DW algae . Total amount of fatty acids varied from 5.5ng fatty acids (FA)/ g POC algae to 16.72ng FA/ g POC al gae according to irradiance of each photobioreactor. Relative amount of un saturated fatty acids (arachidonic acid ARA, eicosapentaenoic acid EPA) to total amount of fatty acids ranged from 4.0 – 8.27% and 9.1 – 18.9%, respec tively. For further details see Chapter 3. The nutrient uptake rates have been determined in 1 hour growth experi ments giving average values of 0.190 – 0.294 mg h 1 PO 4 P per litre culture volume and 0.129 – 0.186 mg h 1 NO 3 N per litre culture volume according to light intensity (see Chapter 3).
The biofilter efficiency of Nannochloropsis sp. were determined with two dif ferent methods: results from one hour growth experiments gave (depending on light intensity) average values for PO 4 P uptake of 0.190 – 0.294 mg h 1 per litre culture volume and for NO 3 N uptake 0.129 – 0.186 mg h 1 per litre culture volume, respectively (see Chapter 3, Tab. 6).
By using these values (average) amount of daily removed nutrients from the recirculation system can be calculated, assuming a total volume of photobio reactors of 3 x 50 litres and 24 hours cultivation time. According to this cal culation the photobioreactor system of the presented dimension can remove a total amount of 0.87g of PO 4 P and 0.57g of NO 3 N per day.
159 Chapter 5
The second method used to calculate the biofilter efficacy of the photobiore actors was the estimation of the daily removal of nutrients by water ex change due to continuous culture mode. According to the adjusted flow rate nutrient rich water from the recirculation system will be introduced at a cer tain volume to the photobioreactors. According to the laws of continuous culture (Pirt, 1975) an equal volume of treated water will flow out of the photobioreactors and back into the main recirculation system at the same time. The difference between the measured dissolved nutrient concentration of the inflow (nutrient rich water) and dissolved nutrient concentration of the outflow (treated water) were determined for each experimental day. Data are shown in Tab. 3. Values vary from 2.1 mg to 28.1 mg PO 4 P per litres dis charge and 0.6 mg to 20.4 mg NO 3 N, because of the variation of the dis solved nutrient concentrations within the photobioreactors during the entire experimental time. However, daily values were multiplied with the total vol ume of recirculation water, passing the photobioreactor system at each ex perimental day according to the adjusted flow rate. Values for total amount of removed PO 4 P and NO 3 N per day vary due to changing conditions within the recirculation system. Surprisingly, taking the average values of all data over the entire experimental period, similar values in comparison to the first calculation were determined: in average 0.88g PO 4 P and 0.57g NO 3 N per day were removed from the recirculation system by the photobioreactor sys tem.
160 Chapter 5
Tab. 3 Differences of the inflow and outflow concentration of the photobioreactor system and daily amount of removed nutrients and flow rates during the experimental period. Optical density (OD 665 ) of the outflow from the harvesting unit. ∆ Date Flow OD 665 inflow – outflow conc. amount of removed nutrients rate per after 1 per day day [L] [mg L ] harvest [g d 1]
PO 4 P NO 3 N TAN NO 2 N PO 4 P NO 3 N TAN NO 2 N 14.10.05 108.0 2.1 3.8 0.23 0.41 15.10.05 108.0 0.062 3.6 5.1 0.39 0.55 19.10.05 108.0 0.138 8.2 18.0 1.14 0.00 0.89 1.94 0.12 0.00 20.10.05 108.0 0.026 4.8 12.5 0.58 0.16 0.52 1.35 0.06 0.02 27.10.05 73.4 0.136 14.5 20.4 1.04 0.02 1.06 1.50 0.08 0.00 28.10.05 73.4 0.102 11.0 26.0 1.38 0.17 0.81 1.9 0.1 0.01 29.10.05 73.4 0.020 14.3 19.5 1.7 0.08 1.05 1.43 0.13 0.01 31.10.05 73.4 0.172 13.8 13.6 0.68 0.08 1.01 1.0 0.05 0.01 01.11.05 73.4 0.002 14.1 5.3 0.1 0.3 1.03 0.39 0.01 0.02 02.11.05 73.4 0.000 15.7 14.0 0.2 0.2 1.15 1.03 0.02 0.02 18.11.05 108.0 0.001 13.9 2.5 0.0 0.1 1.5 0.27 0.00 0.01 23.11.05 86.4 0.001 5.6 1.6 0.1 0.04 0.48 0.14 0.01 0.00 24.11.05 86.4 0.000 14.6 2.8 0.0 0.03 1.26 0.24 0.00 0.00 25.11.05 86.4 0.003 19.7 9.6 0.7 0.01 1.7 0.83 0.06 0.00 27.11.05 86.4 0.002 25.2 2.3 0.0 0.0 2.19 0.20 0.00 0.00 28.11.05 86.4 0.000 25.4 2.2 0.0 0.0 2.19 0.19 0.00 0.00 18.1.06 60.5 0.009 28.1 0.6 1.31 0.02 1.7 0.04 0.08 0.00 19.1.06 14.4 0.013 23.6 2.9 3.19 0.00 0.34 0.04 0.05 0.00 20.1.06 21.6 0.013 23.9 2.9 3.2 0.00 0.52 0.06 0.07 0.00 21.1.06 21.6 0.011 17.6 3.7 2.13 0.06 0.38 0.08 0.05 0.00 22.1.06 21.6 0.012 11.3 3.8 1.94 0.03 0.24 0.08 0.04 0.00 23.1.06 21.6 0.017 12.0 1.7 1.62 0.00 0.26 0.04 0.04 0.00 24.1.06 21.6 0.008 11.0 3.5 0.22 0.07 0.24 0.08 0.01 0.00 4.2.06 14.4 0.025 3.9 0.6 3.8 0.02 0.06 0.01 0.06 0.00 Average 0.88 0.57 0.05 0.00 SD 0.63 0.63 0.04 0.01 5.4 Discussion 5.4.1 Module fish tank The laboratory conditions in MARE II were similar to MARE. However, fish tanks 1 and 2 became too small for the constantly growing fish. In order to analyze the system stability with respect to high total fish biomass, tank 3 (macroalgae tank in MARE) was stocked with additional fish. The load of dis solved nutrients was enhanced for tests of microalgae performance. Due to the changes in the stocking of the tanks suspended solids were transferred mainly to the foam fractionators. At the end of the experimental period the
161 Chapter 5 functionality of foam fractionation was restricted due to large quantities of suspended solids caused by the incomplete first step for removal of sus pended solids. 5.4.2 Module detritivorous tank The performance of the detritivorous reactor as a sink for solid particles pro duced by the fish was sufficient during the growth period of the worms. But, as soon as worm reproduction (monotelic, Nereis diversicolor dies after spawning) started, organic matter was accumulating in the detritivorous tank, resulting in anoxic conditions in turn endangering the successful cul tivation of the fish. The natural life cycle of an organism of critical impor tance for system stability therefore may lead to the breakdown of the system. These results necessarily led to the question of the suitability of Nereis diver- sicolor for integration in aquaculture systems and the search for possible al ternatives. The applicability of sea cucumbers ( Parastichopus californicus, Stychopus japonicus, Ahlgren 1998, Bregman 1994) is discussed as a possi ble alternative organism for Nereis diversicolor in integrated aquacultural systems. The performance of sea cucumbers for removal of settable solids appeared to be good (Ahlgren 1998, Bregman 1994), but the commercial value of these organisms is questionable. Parallel to MARE, the performance of shrimps ( Crangon crangon ) for the removal of settable solids was examined (data not shown). This experiment showed that the performance of Nereis diversicolor can be considered as superior to the shrimps concerning their capacity of settable solid removal. Therefore, currently there seems to be no alternative to Nereis diversicolor as a secondary organism in integrated aqua culture systems. Another possibility of avoiding the problems observed in MARE II is the integration of several detritivorous tanks equipped with Nereis diversicolor in different life cycle stages and their controlled connection to the main system, depending on the stages of life. 5.4.3 Module microalgae bioreactors Although ammonia is the preferred chemical form of nitrogen and readily taken up by phytoplankton (Collos and Slawyk 1981, Levasseur et al. , 1993),
162 Chapter 5 cultivated Nannochloropsis sp. mainly took up nitrate due to the lack of suf ficient ammonia concentrations in the inflow. Nannochloropsis was not sufficiently provided by this nutrient due to the low flow rates through the photobioreactor system. Flow rates are limited by the maximum specific growth rate and may not exceed 0.025 h 1 dilution rate to avoid biomass washout (see Chapter 3). Maximum total water flow per day through the photobioreactors was recorded at 108 L per day, which is still a negligible volume regarding total nutrient budget of the MARE system. Hence, the photobioreactor system can not fulfil the requirements of a biofil ter for removal of toxic ammonia and nitrite from recirculation water. How ever, hints are given that the simultaneous integration with macroalgae filter may be possible, because macroalgae filter are not competing for nitrate due to their preference for ammonia (see Chapter 2). The conceptional design of the photobioreactor system was feasible regarding pre treatment of the water and harvesting process. Cultivated microalgae can be used as valuable feed for feeding organisms ( Brachionus and cope pods), bivalves, Nereis diversicolor or fish larvae (Støttrup and McEvoy, 2002). 5.4.4 Nitrogen cycle Although not primarily intended, this experiment showed substantial in sights into the nitrogen cycle within the system: Over three months, TAN and nitrite concentrations were at low levels. Nitrate concentration was accumulating or remained stable till day 100 indicating nitrification processes. The system started to show instabilities when the onset of worm reproduc tion was observed. The high amount of organic load produced by the fish could not be efficiently degraded by the worms due to their decreased bio mass. Consequently, anaerobic conditions developed in the first sector of the
163 Chapter 5
Nereis-tank, favouring denitrification (Rheinheimer, 1988). Indications for anaerobic conditions were approved during sampling (dark colour of sedi ment, no worms, increased occurrence of gas bubbles). Increasing phosphate values can also be an indication for anaerobic conditions, but in the system, this increase can also be due to the subsequent addition of phosphate via the feed. This addition of phosphate can be considered as the major contri bution to the increasing phosphate values. Evidence for an enhanced denitri fication activity was obtained by i) the rapid decrease of nitrate and ii) the subsequent increase of pH (Rheinheimer, 1988). Additionally, iii) bacterial mats identified as Beggiatoa sp. were observed. Beggiatoa sp. is known as a sulphur oxidising bacterium (Madigan, 2001). The observed black colour of the sediment indicates the presence of H 2S, explaining the presence of Beg- giatoa sp. as an H 2S oxidising bacterium. Furthermore, the ability for deni trification is also described for Beggiatoa sp. on freshwater sediments (Sweerts et al., 1990) and indications were found for denitrification ability of this species in marine habitats as well (McHatton et al., 1996). There is another nitrate consuming process of possible relevance within the system. The reverse course of the nitrification process is called assimilatory nitrate reduction. Many bacterial species are able to reduce nitrate in order to obtain ammonia for biomass synthesis (Madigan, 2001). This process would also lead to a nitrate consumption and the two processes can be con sidered as major processes contributing to the observed decreasing nitrate values. Additionally, nitrate ammonification is performed by several bacteria in order to obtain reduction equivalents for fermentation. However, this process is inhibited by elevated concentrations of ammonia and may there fore be of minor relevance for the explanation of the observed effects. Another process was observed to coincide with decreasing nitrate concentra tions. The TAN concentrations showed strongly increasing values (Fig. 6b and c). Measured TAN values (measured daily) exceeded the TAN concentra tions explicable by fish metabolism (3.35 ± 0.33 mg L 1) only. There are two microbial processes removing ammonia: nitrification and ANAMMOX (Madi gan, 2001). Nitrification is an oxic process and ammonia is thereby con
164 Chapter 5 verted to nitrate. A special trickling biofilter was integrated into the MARE II experiment in order to ensure a constant nitrification process. All surfaces (sediment, tank walls and tubes) are also considered to provide habitats for nitrifying bacteria. Losordo and Wethers (1997) estimated the TAN removal by nitrification outside the biofilter in the range of 30 to 50 %. This could not be proved by Schneider (2000) who showed little contribution of other sys tem components (water, pipes, tanks, sedimentation unit) compared to the aerobic biofilter unit. However during the experiments of Schneider (2000) high values for FCR for African catfish ( Clarias gariepinus ) occurred (experi mental FCR = 2.5). This resulted in high organic waste loads, which Bov endeur (1989) showed will inhibit nitrification processes. Rydl (2005) ob served during her analysis of bacterial activities of a recirculating aquacul ture system that turnover rates of bacteria on wall material and biofilter fill ing material are comparable. Turnover rates of water samples from the sys tem were lower compared to the rates of the wall material and the biofilter filling material. A part of the previously “nitrifying surface” was lost (considering the total nitrification budget) when a part of the sediment surface within the worm tank became anoxic. In this part of the system, ammonia was no longer re moved by nitrification, but nitrate was removed due to denitrification and possible assimilatory nitrate reduction. The increasing nitrite concentrations may also be explained by this process. Nitrite concentrations are supposed to increase, when nitrification is par tially inhibited due to unfavourable environmental conditions (e.g. suboxic/anoxic). ANAMMOX is an anaerobic process converting ammonia and nitrite resulting in the formation of elementary nitrogen. Anoxic conditions were present in the Nereis tank but ammonia as well as nitrite concentrations increased in dicating that the ANAMMOX process was not of major importance in the sys tem.
165 Chapter 5
All these processes are strongly depending on the prevailing oxygen concen trations. Unfortunately, no oxygen concentrations were measured within the different system units, therefore the processes assumed to have happened within the system after the onset of worm reproduction only can rely on measured concentration changes of TAN, NO 2, NO 3. Thus, the cross linking and interactions of these nitrogen converting proc esses, especially nitrification and denitrification strongly depending on each other may explain the observed changes in the system concerning the differ ent nitrogen compounds. The different interactions within the nitrogen cycle had different impacts on the other system modules: i) due to the increased pH, the fraction of dissoci ated ammonia (NH 3) was close to toxic levels (observed = 0.035 mg L 1, litera ture data = 0.05 mg L 1, Losordo et al., 1998). Despite these unfavourable conditions, no mortality and unusual behaviour among the fish could be de tected. Fish even tolerated 10 mg L 1 TAN for a short while at pH 7.96., al though reduced feeding activity occasionally occured. It can therefore be as sumed, that fish is able to adapt to higher ammonia levels, when the water quality is good. ii) the anoxic conditions in the sediment of the worm tank may have led to elevated worm mobility towards the oxic areas within the sediment, in turn increasing the organic load in the first section of the tank and therefore fur ther enhancing denitrification. iii) the foam fractionating process was not efficient enough to successfully remove bacteria attached to particles and microalgae did not show a compa rable performance to the macroalgae for the removal of nitrate also favouring denitrification.
166 Chapter 5
5.4.5 General recommendations The experiments in this study confirmed, that the accurate adaptation of all components of a system is crucial (Losordo et al. , 1999; Waller et al. , 2003). In the first three months of the experiment, MARE II proved to be a stable system, but at the onset of worm reproduction the system started to become instable. Based on the results from this study, the simultaneous use of several biore actors with different worm generations is recommended to allow year round production without any disturbances possibly due to the worm life cycle. It is also recommended, that sediment removed by the worm harvesting process is brought back to the system (assuming aerobic conditions of the Nereis tank). Early life stages of Nereis diversicolor are rather small (< 1mm diame ter) and cannot be easily detected using macroscopic methods. Therefore they are easily lost using the sieving method for harvesting. A mesh size re duction of the sieve would solve this problem, but an efficient sediment re moval is not guaranteed by using smaller mesh sizes, also leading to incor rect worm biomass estimates. The importance of a cautious monitoring of the different nitrogen converting processes is also strongly recommended as a result of this study. Therefore, in possible future experiments it is also crucial to obtain oxygen concentra tions in different compartments of the system and in different regions of the compartments (water, different sediment layers) in order to be able to react before anoxic conditions can develop.
5.5 Acknowledgements This study was funded by Deutsche Bundesstiftung Umwelt (DBU) and the EU (Interreg IIIA). We thank Kerstin Nachtigall for measuring POC and PON.
167 Chapter 5 5.6 References Ahlgren M. O. (1998). Consumption and assimilation of salmon net pen foul ing debris by the Red Sea cucumber Parastichopus californicus : implications for polyculture. Journal of the World Aquaculture Society 29 (2): 133 – 139. Bischoff A.A. (2003). Growth and mortality of the polychaete Nereis diversi- color under experimental rearing conditions. M.Sc.thesis, Institute of Marine Research & Department of Animal Sciences, Chairgroup of Fish Culture and Fisheries, Christian Albrechts Universität Kiel, Germany/Wageningen Uni versity, The Netherlands; 103pp. Bessonart M., Izquierdo M.S., Salhi M., Hernández Cruz C.M., Gonzáles M.M., and Fernández Palacios H. (1999). Effect of dietary arachidonic acid levels on growth and survival of Gilthead seabream. Aquaculture 179: 265 275. Bone Q. and Marshall N. B. (eds.). Biologie der Fische, Gustav Fischer Ver lag, Stuttgart, New York, 1985. Bovendeur, (1989). Fixed Biofilm Reactors applied to Waste Water Treatment and Aquacultural Water Recirculating Systems. Department of Water Pollu tion Control. Wageningen, The Netherlands, Landbouwuniversiteit: 171. Bregman Yu Eh (1994). Bioenergetics of the filter feeding mollusc detritus feeding holothurian food chain under biculture conditions. TINRO, VLADI VOSTOK (Russia). Brown J. A., Wiseman D., and Kean P. (1997). The use of behavioural obser vations in the larviculture of cold water marine fish. Aquaculture 155: 301 – 310. Chopin T., Buschmann A. H., Halling C., Troell M., Kautsky N., Neori A., Kraemer G. P., Zertuche Gonzalez J. A., Yarish C., and Neefus C. (2001). In tegrating seaweeds into marine aquaculture systems: A key toward sustain ability. Journal of Phycology 37 (6): 975 – 986. Chopin T., Bastarache S., Beleyea E., Haya K., Sephton, D., Martin J. L. Eddy S., and Stewart I. (2003). Development of the cultivation of Laminaria saccharina as the extractive inorganic component of an integrated aqucul ture system and monitoring of therapeutants and phycotoxins. Journal of Phycology 39 (S1): 10. Collos Y. and Slawyk G. (1980). Uptake and assimilation by marine phyto plankton. In: Falkowski P.G. Uptake and assimilation by marine phytoplank ton. Plenum Press, New York, pp. 195 211. Hartmann Schröder G. (ed.). Annelida, Borstenwürmer, Polychaeta. Gustav Fischer Verlag, Jena, 1996.
168 Chapter 5
Hussenot J. M. (2003). Emerging effluent management strategies in marine fish culture farms located in European coastal wetlands. Aquaculture 226: 113 – 128. Levasseur M., Thompson P.A. and Harrison P.J. (1993). Physiological accli mation of marine phytoplankton to different nitrogen sources. Journal of Phycology 29: 587 595. Losordo, T. M. and Westers, H. (1997). System Carrying Capacity and Flow Estimations. Aquaculture Water Reuse Systems: Engineering Design and Management. M. B. Timmons and T. M. Losordo. Amsterdam, The Nether lands, Elsevier. 27: 9 – 60. Losordo T.M., Masser M.P., and Rakocy J.E. (1999). Recirculating aquacul ture tank production systems: A review of component options. SRAC Publi cation 453. Madigan M. T., Martinko J. M., and Parker J. (eds) Brock Mikrobiologie. Spektrum Akademischer Verlag Berlin Heidelberg, (2001).
McHatton S . C., Barry J. P., Jannasch H . W., and Nelson D . C. (1996). High Nitrate Concentrations in Vacuolate, Autotrophic Marine Beggiatoa spp.. Appl. Environ. Microbiol. 62 (3): 954 958. Neori A., Shpigel M., and Ben Ezra D. (2000). A sustainable integrated sys tem for culture of fish, seaweed and abalone. Aquaculture 186: 279 – 291. Rheinheimer G., Hegemann W. and Sekoulov R. J. (eds.). Stickstoffkreislauf im Wasser: Stickstoffumsetzung in natürlichen Gewässern, in der Abwasser reinigung und Wasserversorgung. Oldenburg Verlag, München, 1988. Rydl, A. (2005). Sukzession und Analyse des bakteriellen Bewuchses an Fi schen und Materialoberflächen in einer geschlossenen Fischzuchtanlage. Diplomarbeit, Leibniz Institute of Marine Sciences, Kiel., 103 pp. Sargent J.R., McEvoy L.A., and Bell J.G. (1997). Requirements, presentation and sources of polyunsaturated fatty acids in marine fish larval feeds. Aqua culture 155: 117 127. Schneider, O. (2000). Modelling Aquaculture Systems: Energy flow and nu trient flows in commercial catfish farms, using recirculation systems. M. Sc. thesis, Fish Culture and Fisheries Group, Wageningen University and the Institute for Marine Research, University of Kiel, 123pp. Shpigel M., Neori A., Popper D. M., and Gordin H. (1993). A proposed model for “environmentally clean” land based culture of fish, bivalves and sea weeds. Aquaculture 117: 115 – 128. Støttrup J.G. and McEvoy L.A. (eds.) (2002). Live feeds in marine aquacul ture. Blackwell publishing 336pp.
169 Chapter 5
Sweerts J. P. R. A., De Beer D., Nielsen L. P., Verdouw H., Van den Heuvel J. C., Cohen Y., and Cappenberg T. E. (1990). Denitrification by sulphur oxidizing Beggiatoa spp. mats on freshwater sediments. Nature, Vol ume 344, Issue 6268: 762 763. Tandler A., Mozes N., and Ucko M. (2003). The assimilation of dissolved fish waste by microalgae. Annual report 2003 of the EU project ZAFIRA (Zero discharge Aquaculture by Farming in Integrated Recirculation Systems in Asia), http://zafira.wau.nl: 42 69. Waller U. (2000). Tank culture – including raceways and re circulating sys tems. In: Environmental impacts of aquaculture. Edited by K.D.Black. Shef field Academic Press. Waller U., Sander M., and Piker L. (2001). Low energy and low water con sumption recirculation system for marine fish: first results from a test run with Dicentrarchus labrax in an improved recirculating system and sugges tions on an integration into secondary production lines. European Aquacul ture Society Special Publications 29: 265 266. Waller U., Bischoff A.A., Orellana J., Sander M., and Wecker B. (2003). An advanced technology for clear water aquaculture recirculation systems: Re sults from a pilot production of Sea bass and hints towards "Zero Dis charge". European Aquaculture Society Special Publications 33: 356 357.
170
Danksagung Ich danke Frau Prof. Dr. Karin Lochte für ihre Bereitschaft, diese Arbeit zu betreuen. Ihr Vertrauen, Einsatz und ihre Diskussionsbereitschaft waren die Grundlage dieser Arbeit. Herrn Prof. Dr. Schnack danke ich für die Unterstützung während der Pro motionszeit und Herrn Prof. Dr. Dr. h.c. Harald Rosenthal für seine Hilfe und seinen Einsatz bei der Fertigstellung dieser Arbeit. Meinen Kollegen Adrian und Bert danke ich für die gemeinsame Zeit, für ihre Freundschaft und für jede einzelne Stunde, die wir an unseren Anlagen und im Labor verbracht haben! Uwe Waller hat als Leiter der Gruppe für die nöti ge Finanzierung und technischen Voraussetzungen gesorgt. Er war als Dis kussionspartner, Rat und Ideengeber immer eine große Hilfe. Großen Anteil am Gelingen dieser Arbeit haben Michael, Ralf und alle ande ren vom Aquarium. Ihr wart die feste Basis dieser Arbeit, eure Ideen und das besondere persönliche Verhältnis waren sehr wichtig. Besonders wohl habe ich mich immer bei den Mikrobiologen gefühlt. Bei der Gruppe von Herrn Prof. Hans Georg Hoppe in der Hohenbergstr. habe ich sehr viel Zeit verbracht und genoss die angenehme Atmosphäre. Im Beson deren möchte ich mich bei Regine Koppe bedanken, ohne ihre Hilfe wären die mikrobiologischen Untersuchungen nicht möglich gewesen. Aber auch die Gruppe im Haupthaus (Jutta Wiese, Vera Thiel und all die anderen) hieß mich immer willkommen und stand mit Rat und Tat (und Autoklaven) zur Seite. Line Christensen und Jens Jorgen Lonsman Iversen waren wertvolle Pro jektpartner, die mit ihrem Wissen über Bioreaktorentechnik dieses Projekt erst möglich gemacht haben. Dr. Patrick Fink vom MPI Plön danke ich für die unkomplizierte Zusammen arbeit bei der Analyse der Fettsäureprofile. Besonderer Dank gilt dem technischen Personal des IFM GEOMAR: Kerstin Nachtigall, Peter Fritsche und Thomas Hansen haben durch ihre Analysen einen wichtigen Beitrag zu dieser Arbeit geleistet. Ganz besonders möchte ich Helge Mempel danken, der immer geholfen hat, wenn der Laboralltag seine Tücken hatte. Hans Langmaack und Dirk Wehrend von der Zentral werkstatt waren immer verfügbar und haben so manche Sonderanfertigung realisiert. Günther Peters war eine große Unterstützung bei kleinen und gro ßen Problemen in der elektrischen Versorgung der Kreislaufanlagen. Danke an alle, die hier nicht namentlich aufgeführt werden konnten! Die Abteilung Fischereibiologie mit all ihren Doktoranden, Postdocs und technischen Mitarbeitern hat während der gesamten Zeit eine schöne und angenehme Arbeitsatmosphäre geschaffen. Es hat mich gefreut, ein Teil da von gewesen zu sein. Jan, danke für Deine Hilfe in Ozonfragen und Jörn für
die Unterstützung bei Computerproblemen. Helmut ist während der gesam ten Zeit eine wichtige Unterstützung gewesen und hat sehr viel Geduld bei den teilweise belagerungsartigen Zuständen im Büro bewiesen. Ich danke meinen Eltern, dass sie mir ermöglicht haben, meinen Traum zu erfüllen und immer hinter mir gestanden haben; meinen Verwandten für das Verständnis ob der knappen Zeit für gemeinsame Stunden. Marco, Dir danke ich für Deine Hilfe, Dein Verständnis und die Fähigkeit, mich immer wieder aufzurichten. Auch meinen Freunden, allen voran Elke, Flo, Lenard, Katja und Steffi habe ich es zu verdanken, dass ich diese Arbeit beenden konnte. Tini, Dir gebührt mein allergrößter Dank für deinen spontanen, selbstlosen und ermüdlichen Einsatz in der doch sehr turbulenten Endphase meiner Arbeit. Deine Anregungen und Korrekturen haben der Arbeit den nötigen Schliff gegeben und dein Optimismus haben noch einmal alle Kräfte in mir mobilisiert. Ich werde Dir und Adrian nie vergessen, dass Ihr mich in der schwersten Zeit nicht allein gelassen habt.
Lebenslauf
Persönliche Daten Name: Kube Vorname: Nicole Geburtsdatum: 30. September 1976 Geburtsort: Burg (bei Magdeburg) Staatsangehörigkeit: deutsch Schulische Ausbildung 1983 – 1991 Besuch der Polytechnischen Oberschulen „Pestalozzi“ und „W.I. Lenin“, Burg 1991 – 1995 Gymnasium Burg, Abschluss mit Abitur
Studium
1995 2001 Studium der Biologie an der Christian Albrechts Universität zu Kiel, mit den Schwerpunkten Zoologie und Fischereibiologie, Nebenfächer: Meereschemie Berufliche Tätigkeiten 2000 2003 Journalistische Ausbildung und freie Mitarbeitertätigkeit bei der Filmproduktionsfirma BLUE PLANET FILM in Hamburg 2003 Erstellung der Studie „Marine Naturstoffe in der Blauen Biotechnologie: Stand und Perspektiven“ im Auftrag der Innovationsstiftung Schleswig Holstein
Aug. 2003 – Mai 2006 Wissenschaftliche Mitarbeiterin am Leibniz Institut für Meereswissenschaften Kiel Promotionsthema: “The integration of microalgae photobio reactors in a recirculation system for low water discharge mariculture“
Erklärung
Hiermit erkläre ich, dass die vorliegende Arbeit nach Inhalt und Form und die ihr zugrunde liegenden Versuche meine eigene Arbeit sind. Es wurden – abgesehen von der Beratung durch meine akademischen Lehrer – keine an deren als die angegebenen Hilfsmittel und Quellen verwendet. Wörtlich und inhaltlich aus anderen Quellen entnommene Textstellen sind als solche kenntlich gemacht. Diese Arbeit wurde weder ganz noch in Auszügen an einer anderen Stelle im Rahmen eines Prüfungsverfahrens vorgelegt. Ferner erkläre ich hiermit, dass ich noch keine früheren Promotionsversuche unternommen habe. Für die Prüfung wird die Form der Disputation gewählt. Der Zulassung von Zuhörern/Zuhörerinnen bei der mündlichen Prüfung wird nicht widerspro chen. Kiel, den ______ Nicole Kube